TW201909643A - Image processing method, device and system - Google Patents
Image processing method, device and system Download PDFInfo
- Publication number
- TW201909643A TW201909643A TW107116752A TW107116752A TW201909643A TW 201909643 A TW201909643 A TW 201909643A TW 107116752 A TW107116752 A TW 107116752A TW 107116752 A TW107116752 A TW 107116752A TW 201909643 A TW201909643 A TW 201909643A
- Authority
- TW
- Taiwan
- Prior art keywords
- image block
- current
- reconstructed image
- filter
- reconstructed
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/132—Sampling, masking or truncation of coding units, e.g. adaptive resampling, frame skipping, frame interpolation or high-frequency transform coefficient masking
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/117—Filters, e.g. for pre-processing or post-processing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/44—Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/70—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
本發明涉及圖像處理技術,尤其涉及一種圖像處理方法、設備及系統。The invention relates to image processing technology, in particular to an image processing method, device and system.
數位影像是以數位形式記錄的影像。圖1為本發明提供的數位影像的示意圖,如圖1所示,數位影像由多序框數位圖像組成。圖2為本發明提供的數位圖像的示意圖,如圖2所示,圖像由12*16個畫素組成,其中,每個畫素被稱為一個畫素,12*16表示圖像解析度。例如2K影像的圖像解析度是1920*1080,4K影像的圖像解析度是3840*2160。通常原始影像包括的資料量較大,不適合儲存和傳輸,需要利用高效的影像壓縮編碼技術來壓縮原始資料。Digital images are images recorded in digital form. FIG. 1 is a schematic diagram of a digital image provided by the present invention. As shown in FIG. 1, the digital image is composed of multi-order frame digital images. FIG. 2 is a schematic diagram of a digital image provided by the present invention. As shown in FIG. 2, the image is composed of 12 * 16 pixels, where each pixel is called a pixel, and 12 * 16 represents image analysis degree. For example, the image resolution of 2K video is 1920 * 1080, and the image resolution of 4K video is 3840 * 2160. Generally, the original image contains a large amount of data, which is not suitable for storage and transmission. It is necessary to use efficient image compression coding technology to compress the original data.
具體地,圖3為本發明提供的編碼端的編碼示意圖,如圖3所示,編碼端的編碼流程包括:編碼端接收到影像之後,對於構成影像的每序框圖像,將該圖像劃分成多個待編碼圖像區塊。對於當前待編碼圖像區塊,首先通過參考重建圖像區塊(該參考重建圖像區塊用於提供當前待編碼圖像區塊所需的參考畫素,該參考畫素用於對當前待編碼圖像區塊進行預測)對當前待編碼圖像區塊進行預測,得到當前待編碼圖像區塊的預測訊號;用當前待編碼圖像區塊的原始訊號減去預測訊號,得到殘差訊號。經過預測後,殘差訊號的振幅遠小於原始訊號。將殘差訊號進行變換和量化操作。經過變換量化後,得到變換量化係數,再通過亂度編碼技術編碼量化係數以及編碼中的其他指示資訊,得到位元串流。進一步地,編碼端還需要重建當前待編碼圖像區塊,以實現對後續待編碼圖像區塊的編碼提供參考畫素。具體地,在得到當前待編碼圖像區塊的變換量化係數之後,編碼端需要對當前待編碼圖像區塊的變換量化係數進行反量化和反變換,得到重建的殘差訊號,將重建的殘差訊號與當前待編碼圖像區塊對應的預測訊號相加,得到當前待編碼圖像區塊的重建訊號,根據該重建訊號得到重建圖像區塊。其中,該重建圖像區塊可以對後續待編碼圖像區塊進行預測。可選地,殘差訊號經過變換後得到變換係數,變換係數通過量化後會有資訊損失,該資訊損失不可逆。即經過反量化後的變換係數會有失真,從而使得重建訊號與原始訊號不一致,這種壓縮方式為失真壓縮。因此,對於失真壓縮,在得到重建圖像區塊後,需要對該重建圖像區塊進行濾波,從而去除失真壓縮引入的一些失真,例如方塊效應,漣波效應等。為了去除方塊效應,可以使用H.264、H.265標準中的DBK濾波器。為了去除漣波效應,可以使用H.265中的SAO濾波器,以及下一代標準中的ALF濾波器等。也有無失真壓縮方法,即殘差訊號使用無損的變換操作得到變換係數,不進行量化操作,將變換係數進行亂度編。對於無失真壓縮,一般不再進行濾波操作。進一步地,當前圖像的各個圖像區塊都完成重建之後,得到重建圖像,其中,該重建圖像可以對後續其他序框圖像進行預測。Specifically, FIG. 3 is a schematic diagram of the encoding end provided by the present invention. As shown in FIG. 3, the encoding process of the encoding end includes: after the encoding end receives the image, for each sequence frame image constituting the image, the image is divided into Multiple image blocks to be encoded. For the current image block to be encoded, first reconstruct the image block by reference (the reference reconstruction image block is used to provide the reference pixel required by the current image block to be encoded, and the reference pixel is used to Predict the image block to be encoded) predict the current image block to be encoded to obtain the prediction signal of the current image block to be encoded; subtract the prediction signal from the original signal of the current image block to be encoded to obtain the residual signal Bad signal. After prediction, the amplitude of the residual signal is much smaller than the original signal. Transform and quantize the residual signal. After transform and quantization, transform and quantization coefficients are obtained, and then the quantization coefficients and other indication information in the encoding are encoded by the random coding technology to obtain a bit stream. Further, the encoding end also needs to reconstruct the current image block to be encoded, so as to provide reference pixels for encoding the subsequent image block to be encoded. Specifically, after obtaining the transform and quantization coefficients of the current image block to be encoded, the encoding end needs to inversely quantize and inverse transform the transform and quantization coefficients of the current image block to obtain the reconstructed residual signal, and The residual signal is added to the prediction signal corresponding to the current image block to be encoded to obtain the reconstruction signal of the current image block to be encoded, and the reconstructed image block is obtained according to the reconstruction signal. Among them, the reconstructed image block can predict the subsequent image block to be encoded. Optionally, the transform coefficients are obtained after the residual signal is transformed, and there will be information loss after the transform coefficients are quantized, and the information loss is irreversible. That is, the transform coefficient after inverse quantization will be distorted, so that the reconstructed signal is inconsistent with the original signal. This compression method is distortion compression. Therefore, for distortion compression, after the reconstructed image block is obtained, it is necessary to filter the reconstructed image block, so as to remove some distortions introduced by the distortion compression, such as the square effect and the ripple effect. In order to remove the square effect, the DBK filter in the H.264 and H.265 standards can be used. In order to remove the ripple effect, you can use the SAO filter in H.265, and the ALF filter in the next generation standard. There is also a distortionless compression method, that is, the residual signal uses a lossless transform operation to obtain transform coefficients, and the quantization operation is not performed, and the transform coefficients are scrambled. For distortionless compression, the filtering operation is generally not performed. Further, after each image block of the current image has been reconstructed, a reconstructed image is obtained, wherein the reconstructed image can be used to predict other subsequent frame images.
圖4為本發明提供的解碼端的解碼示意圖,如圖4所示,解碼端獲取到位元串流之後,首先對位元串流進行亂度解碼,得到當前待重建圖像區塊的變換量化係數,然後對變換量化係數進行反量化和反變換,得到當前待重建圖像區塊的重建的殘差訊號。通過它的參考重建圖像區塊對當前待重建圖像區塊進行預測,得到當前待重建圖像區塊的預測訊號,然後將預測訊號和上述重建的殘差訊號相加,得到當前待重建圖像區塊的重建訊號,然後根據該重建訊號得到當前待重建圖像區塊對應的當前重建圖像區塊,其中,該當前重建圖像區塊可以對後續其他待重建圖像區塊進行預測。類似於上述編碼端的情況,可選地,在解碼端需要對當前重建圖像區塊進行濾波。進一步地,當前圖像的各個圖像區塊都完成重建之後,得到重建圖像,其中,該重建圖像可以對後續其他序框圖像進行預測。FIG. 4 is a schematic diagram of decoding at a decoding end provided by the present invention. As shown in FIG. 4, after obtaining the bit stream, the decoding end first performs random decoding on the bit stream to obtain transform and quantization coefficients of the current image block to be reconstructed Then, inverse quantization and inverse transformation are performed on the transformed quantization coefficients to obtain the reconstructed residual signal of the current image block to be reconstructed. Predict the current image block to be reconstructed through its reference reconstruction image block to obtain the prediction signal of the current image block to be reconstructed, and then add the prediction signal and the reconstructed residual signal to obtain the current reconstruction target The reconstruction signal of the image block, and then obtain the current reconstruction image block corresponding to the current reconstruction image block according to the reconstruction signal, wherein the current reconstruction image block can be performed on the subsequent other reconstruction image blocks prediction. Similar to the situation at the encoding end, optionally, the decoding end needs to filter the current reconstructed image block. Further, after each image block of the current image has been reconstructed, a reconstructed image is obtained, wherein the reconstructed image can be used to predict other subsequent frame images.
為了降低編解碼複雜度,編碼端對每序框圖像先進行縮減取樣處理,圖5為本發明提供的編碼端的編碼示意圖,如圖5所示,編碼端對整幅圖像採用縮減取樣處理,然後對經過縮減取樣處理後的圖像中的每個待編碼圖像區塊進行編碼,得到位元串流。其中,每個待編碼圖像區塊對應的重建圖像區塊的解析度為縮減取樣解析度。相應的,解碼端解析位元串流,每個待重建圖像區塊的解析度均為縮減取樣解析度,得到對應的重建圖像區塊的解析度也為縮減取樣解析度,解碼端需要對重建圖像區塊採用升取樣處理,以得到原始解析度的重建圖像區塊。In order to reduce the complexity of encoding and decoding, the encoding end performs down-sampling processing on each sequence frame image. FIG. 5 is a schematic diagram of the encoding end provided by the present invention. As shown in FIG. 5, the encoding end uses down-sampling processing for the entire image Then, encode each image block to be encoded in the image after downsampling to obtain a bit stream. The resolution of the reconstructed image block corresponding to each image block to be encoded is the down-sampling resolution. Correspondingly, the decoding end parses the bit stream. The resolution of each image block to be reconstructed is the down-sampling resolution. The resolution of the corresponding reconstructed image block is also the down-sampling resolution. The decoder needs to Upsampling is applied to the reconstructed image block to obtain the reconstructed image block with original resolution.
現有技術中,編碼端或者解碼端對於整幅圖像中每個重建圖像區塊均採用相同的濾波器進行升取樣處理,然而,各個重建圖像區塊的特性可能不同,例如,有些重建圖像區塊可能較為平坦;有些重建圖像區塊可能細節比較多,現有技術勢必導致有些經過升取樣處理之後的重建圖像區塊比較模糊,存在顯示效果不佳的問題。In the prior art, the encoding end or the decoding end uses the same filter for upsampling processing for each reconstructed image block in the entire image, however, the characteristics of each reconstructed image block may be different, for example, some reconstructions The image blocks may be relatively flat; some reconstructed image blocks may have more details. The prior art will inevitably cause some reconstructed image blocks that have undergone upsampling to be blurred and have a problem of poor display effect.
本發明提供圖像處理方法、設備及系統,從而解決有些經過升取樣處理之後的重建圖像區塊顯示效果不佳的問題。The invention provides an image processing method, equipment and system, so as to solve the problem of poor display effect of some reconstructed image blocks after upsampling processing.
第一方面,本發明提供一種圖像處理方法,包括:生成當前待編碼圖像區塊的重建訊號,並根據重建訊號重建當前待編碼圖像區塊,以得到當前重建圖像區塊;若當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,並通過第一濾波器對當前重建圖像區塊進行升取樣處理。In a first aspect, the present invention provides an image processing method, including: generating a reconstruction signal of a current image block to be encoded, and reconstructing the current image block to be encoded according to the reconstruction signal to obtain the current reconstruction image block; if The coding method of the current reconstructed image block is the down-sampling coding method, then the first filter for upsampling the current reconstructed image block is selected from at least two candidate filters, and passes through the first filter Upsampling the current reconstructed image block.
本發明的功效為:相對於現有技術中對整幅圖像中的重建圖像區塊都採用相同的濾波器,本發明針對每個重建圖像區塊都要選擇對應的濾波器,即本發明採用具有針對性的選擇方式來選擇濾波器,通過選擇的濾波器對重建圖像區塊進行升取樣處理,從而可以得到顯示效果更佳的重建圖像區塊。The effect of the present invention is that, compared with the prior art, which uses the same filter for the reconstructed image blocks in the entire image, the present invention selects the corresponding filter for each reconstructed image block, that is, the present The invention adopts a targeted selection method to select a filter, and performs up-sampling processing on the reconstructed image block through the selected filter, so that a reconstructed image block with better display effect can be obtained.
可選地,從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,具體包括:根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器中選擇第一濾波器。Optionally, selecting a first filter for upsampling the current reconstructed image block from at least two candidate filters, specifically including: selecting from at least two candidates according to the texture characteristics of the current reconstructed image block Select the first filter among the filters.
可選地,根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器中選擇第一濾波器,包括:根據預設映射關係和所述當前重建圖像區塊的紋理特徵選擇第一濾波器,預設映射關係是包括當前重建圖像區塊的紋理特徵的預設紋理特徵和包括第一濾波器的至少兩個候選濾波器之間的映射關係。Optionally, selecting the first filter from at least two candidate filters according to the texture feature of the current reconstructed image block includes: selecting the first filter according to a preset mapping relationship and the texture feature of the current reconstructed image block For the filter, the preset mapping relationship is a mapping relationship between the preset texture feature including the texture feature of the currently reconstructed image block and at least two candidate filters including the first filter.
通過根據當前重建圖像區塊的紋理特徵為當前重建圖像區塊選擇第一濾波器,通過選擇的濾波器對重建圖像區塊進行升取樣處理,從而可以得到顯示效果更佳的重建圖像區塊。By selecting the first filter for the current reconstructed image block according to the texture features of the current reconstructed image block, and performing upsampling processing on the reconstructed image block through the selected filter, a better displayed reconstruction image Like a block.
可選地,從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,具體包括:確定當前重建圖像區塊的至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊與當前重建圖像區塊的相似度,其中,至少兩個相鄰重建圖像區塊進行升取樣處理時對應至少兩個第二濾波器,至少兩個候選濾波器包括至少兩個第二濾波器;從至少兩個第二濾波器中選擇與當前重建圖像區塊相似度最高的相鄰重建圖像區塊對應的第二濾波器作為第一濾波器。Optionally, selecting a first filter for upsampling the current reconstructed image block from at least two candidate filters specifically includes: determining at least two adjacent reconstruction images of the current reconstructed image block The similarity between each adjacent reconstructed image block in the image block and the current reconstructed image block, where at least two adjacent reconstructed image blocks correspond to at least two second filters when upsampling is performed, The at least two candidate filters include at least two second filters; from the at least two second filters, the second filter corresponding to the adjacent reconstructed image block with the highest similarity of the current reconstructed image block is selected as the second filter First filter.
通過根據當前重建圖像區塊的相鄰重建圖像區塊選擇第一濾波器,通過選擇的濾波器對重建圖像區塊進行升取樣處理,從而可以得到顯示效果更佳的重建圖像區塊。By selecting the first filter according to the adjacent reconstructed image block of the current reconstructed image block, and upsampling the reconstructed image block through the selected filter, a reconstructed image area with better display effect can be obtained Piece.
可選地,從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器包括,包括:通過至少兩個候選濾波器分別對所述當前重建圖像區塊進行升取樣處理,得到至少兩個候選濾波器分別對應的升取樣圖像區塊;分別計算至少兩個候選濾波器對應的升取樣圖像區塊與所述當前重建圖像區塊對應的原始圖像區塊的誤差;將最小誤差對應的候選濾波器作為第一濾波器。通過該方法可以更加精確的選擇第一濾波器。Optionally, selecting the first filter for upsampling the current reconstructed image block from at least two candidate filters includes: separately performing the current reconstructed image through at least two candidate filters The block is up-sampled to obtain up-sampled image blocks corresponding to at least two candidate filters respectively; the up-sampled image blocks corresponding to at least two candidate filters are calculated to correspond to the current reconstructed image blocks The error of the original image block; the candidate filter corresponding to the smallest error is used as the first filter. By this method, the first filter can be selected more accurately.
可選地,通過第一濾波器對所述當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對所述當前重建圖像區塊進行一次升取樣處理;相應的,方法還包括:若所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: according to the current value of the adjacent reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of a part of the adjacent reconstructed image blocks that have been reconstructed are subjected to an up-sampling process of the current reconstructed image block through the first filter; correspondingly, the method further includes: if necessary adjacent reconstruction In the image block, another part of the neighboring reconstructed image block that has not been reconstructed currently has been reconstructed, and then passes the third filter to the part of the current reconstructed image block according to another part of the neighboring reconstructed image block. The boundary is subjected to double-upsampling processing, in which part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
可選地,通過第一濾波器對當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對當前重建圖像區塊進行一次升取樣處理;相應的,所述方法還包括:若當前重建圖像區塊所處的當前圖像的所有圖像區塊均已完成重建,則通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,另一部分相鄰重建圖像區塊是在對當前重建圖像區塊進行第一次升取樣處理時未完成重建的圖像區塊;當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: the currently completed neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block A part of the pixels of the adjacent reconstructed image block that are reconstructed are subjected to an up-sampling process on the current reconstructed image block through the first filter; correspondingly, the method further includes: if the current reconstructed image block is located All the image blocks of the current image have been reconstructed, then pass the third filter to the current reconstruction image according to another part of the adjacent reconstructed image blocks required Perform partial upsampling on part of the boundary of the image block, where another part of the adjacent reconstructed image block is the image block that has not been reconstructed when the first reconstructed image block is first upsampled; Part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
本發明這兩種可選方法可以避免當前重建圖像區塊邊界不連續的問題。The two optional methods of the present invention can avoid the problem of discontinuous boundary of the current reconstructed image block.
可選地,第三濾波器為所述第一濾波器。Optionally, the third filter is the first filter.
可選地,通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理之前,還包括:根據另一部分相鄰重建圖像區塊和當前重建圖像區塊的部分邊界判斷是否對部分邊界進行二次升取樣處理;若確定對第一重建圖像區塊的部分邊界進行二次升取樣處理,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, before the second upsampling process is performed on a part of the boundary of the current reconstructed image block according to another part of the neighboring reconstructed image block required by the third filter through the third filter, It also includes: judging whether to perform the sub-upsampling process on the partial boundary according to another part of the adjacent reconstructed image block and the partial boundary of the current reconstructed image block; For the sub-upsampling process, a third filter is used to perform a sub-upsampling process on part of the boundary of the current reconstructed image block according to another part of the adjacent reconstructed image block.
當確定不對當前重建圖像區塊進行二次升取樣處理時,可以降低解碼端的開銷,當確定需要對當前重建圖像區塊進行二次升取樣處理時,可以解決當前重建圖像區塊的邊界不連續問題。When it is determined that the current reconstructed image block is not to be upsampled, the overhead of the decoding end can be reduced. When it is determined that the current reconstructed image block needs to be upsampled, the current reconstructed image block can be resolved The problem of boundary discontinuity.
可選地,還包括:生成位元串流,該位元串流包括:第一濾波器的標識資訊。Optionally, the method further includes: generating a bit stream, the bit stream includes: identification information of the first filter.
可選地,位元串流還包括第一指示資訊,其中,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器。Optionally, the bit stream further includes first indication information, where the first indication information is used to indicate how to select a filter used for upsampling the current reconstructed image block from at least two candidate filters.
可選地,生成位元串流,位元串流包括:第二指示資訊,其中,第二指示資訊用於指示解碼端是否需要對當前重建圖像區塊進行二次升取樣處理。Optionally, a bit stream is generated. The bit stream includes: second indication information, where the second indication information is used to indicate whether the decoding end needs to perform up-sampling processing on the current reconstructed image block.
下面將介紹解碼端的圖像處理方法,其效果與編碼端對應效果類似,下面不再贅述。The image processing method at the decoding end will be described below, and its effect is similar to the corresponding effect at the encoding end, which will not be repeated below.
第二方面,本發明提供一種圖像處理方法,包括:解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和所述當前待重建圖像區塊的編碼方式;根據當前待重建圖像區塊的編碼資訊生成當前待重建圖像區塊的重建訊號,並根據重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則根據從位元串流獲取的第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,其中,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;通過第一濾波器對當前重建圖像區塊進行升取樣處理。In a second aspect, the present invention provides an image processing method, including: parsing a bit stream to obtain encoding information of a current image block to be reconstructed and the encoding method of the current image block to be reconstructed; The encoded information of the reconstructed image block generates the reconstruction signal of the current image block to be reconstructed, and reconstructs the current image block to be reconstructed according to the reconstruction signal to obtain the current reconstruction image block; if the current image block to be reconstructed The encoding method is a down-sampling encoding method, and the first filter used for upsampling the current reconstructed image block is selected according to the first indication information obtained from the bit stream, where the first indication information is used to indicate How to select the filter used for upsampling the current reconstructed image block from at least two candidate filters; upsampling the current reconstructed image block through the first filter.
可選地,第一指示資訊用於指示根據當前重建圖像區塊的紋理特徵選擇濾波器從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;根據從所述位元串流獲取的第一指示資訊選擇用於對所述當前重建圖像區塊進行升取樣處理的第一濾波器,包括:根據第一指示資訊確定當前重建圖像區塊的紋理特徵;根據預設映射關係和當前重建圖像區塊的紋理特徵選擇第一濾波器,預設映射關係是包括當前重建圖像區塊的紋理特徵的預設紋理特徵和包括第一濾波器的所述至少兩個候選濾波器之間的映射關係。Optionally, the first indication information is used to instruct the filter selection according to the texture feature of the current reconstructed image block to select the filter used for upsampling the current reconstructed image block from at least two candidate filters; The first indication information acquired from the bit stream selects a first filter for upsampling the current reconstructed image block, including: determining the current reconstruction image block according to the first indication information Texture feature; the first filter is selected according to the preset mapping relationship and the texture feature of the currently reconstructed image block, the preset mapping relationship is the preset texture feature including the texture feature of the currently reconstructed image block and includes the first filter The mapping relationship between the at least two candidate filters.
可選地,第一指示資訊用於指示根據當前重建圖像區塊的相鄰重建圖像區塊從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;根據從位元串流獲取的第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,包括:確定至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊與當前重建圖像區塊的相似度,至少兩個相鄰重建圖像區塊進行升取樣處理時對應至少兩個第二濾波器,至少兩個候選濾波器包括至少兩個第二濾波器;從至少兩個第二濾波器中選擇相似度最高的相鄰重建圖像區塊對應的第二濾波器作為所述第一濾波器。Optionally, the first indication information is used to instruct the filter used for upsampling the current reconstructed image block to be selected from at least two candidate filters according to adjacent reconstructed image blocks of the current reconstructed image block Selecting the first filter for upsampling the current reconstructed image block according to the first indication information obtained from the bit stream, including: determining each phase in at least two adjacent reconstructed image blocks The similarity between the neighboring reconstructed image block and the current reconstructed image block. At least two adjacent reconstructed image blocks correspond to at least two second filters when upsampling is performed. The at least two candidate filters include at least two A second filter; a second filter corresponding to the adjacent reconstructed image block with the highest similarity from at least two second filters is selected as the first filter.
可選地,通過第一濾波器對當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對當前重建圖像區塊進行一次升取樣處理;相應的,方法還包括:若所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,當前重建圖像區塊的所述部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: the currently completed neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of a part of the reconstructed adjacent reconstructed image blocks are subjected to an up-sampling process of the current reconstructed image block through the first filter; correspondingly, the method further includes: if necessary adjacent reconstructed image blocks In the currently uncompleted reconstruction of another part of the adjacent reconstructed image block in which reconstruction has been completed, a third filter is used to perform partial boundary reconstruction of the current reconstructed image block according to another part of the adjacent reconstructed image block Double upsampling processing, wherein the part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
可選地,通過第一濾波器對當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對當前重建圖像區塊進行一次升取樣處理;相應的,方法還包括:若當前重建圖像區塊所處的當前圖像的所有圖像區塊均已完成重建,則通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,另一部分相鄰重建圖像區塊是在對當前重建圖像區塊進行第一次升取樣處理時未完成重建的圖像區塊;當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: the currently completed neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of a part of the reconstructed adjacent reconstructed image blocks are subjected to an up-sampling process of the current reconstructed image block through the first filter; correspondingly, the method further includes: if the current reconstructed image block is located After all image blocks of the image have been reconstructed, the third filter is used to determine the current reconstructed image area according to another part of the neighboring reconstructed image blocks required. The part of the boundary of the block is subjected to double upsampling, where another part of the adjacent reconstructed image block is an image block that has not been reconstructed when the first upsampling process is performed on the current reconstructed image block; the current reconstruction Part of the boundary of the image block is adjacent to another part of the adjacent reconstructed image block.
可選地,所述第三濾波器為所述第一濾波器。Optionally, the third filter is the first filter.
可選地,通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理之前,還包括:根據另一部分相鄰重建圖像區塊和當前重建圖像區塊的部分邊界判斷是否對部分邊界進行二次升取樣處理;若確定對第一重建圖像區塊的部分邊界進行二次升取樣處理,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, before the second upsampling process is performed on a part of the boundary of the current reconstructed image block according to another part of the neighboring reconstructed image block required by the third filter through the third filter, It also includes: judging whether to perform the sub-upsampling process on the partial boundary according to another part of the adjacent reconstructed image block and the partial boundary of the current reconstructed image block; For the sub-upsampling process, a third filter is used to perform a sub-upsampling process on part of the boundary of the current reconstructed image block according to another part of the adjacent reconstructed image block.
可選地,位元串流還包括:第二指示資訊;相應的,通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,包括:若第二指示資訊指示需要對當前重建圖像區塊進行二次升取樣處理,則通過第三濾波器,根據在所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, the bit stream further includes: second indication information; correspondingly, a third filter is used to reconstruct the current reconstruction image block according to another part of the adjacent reconstruction image blocks required Performing sub-upsampling processing on part of the boundary of the image block, including: if the second indication information indicates that it is necessary to perform up-sampling processing on the current reconstructed image block, then pass a third filter according to the required phase Another part of the neighboring reconstructed image block in the neighboring reconstructed image block performs sub-upsampling processing on a part of the boundary of the current reconstructed image block.
第三方面,本發明提供一種圖像處理方法,包括:解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和當前待重建圖像區塊的編碼方式;根據編碼資訊生成當前待重建圖像區塊的重建訊號,並根據重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則從位元串流獲取第一濾波器的標識資訊,並通過標識資訊標識的第一濾波器對當前重建圖像區塊進行升取樣處理。In a third aspect, the present invention provides an image processing method, including: parsing a bit stream to obtain the encoding information of the current image block to be reconstructed and the encoding method of the current image block to be reconstructed; generating the current according to the encoding information The reconstruction signal of the image block to be reconstructed, and reconstruct the current image block to be reconstructed according to the reconstruction signal to obtain the current reconstruction image block; if the current encoding method of the image block to be reconstructed is a down-sampling encoding method, then from The bit stream obtains the identification information of the first filter, and upsamples the current reconstructed image block through the first filter identified by the identification information.
下面對圖像處理設備及系統進行介紹,其實現原理和技術效果與上述原理和技術效果類似,此處不再贅述。The following describes the image processing equipment and system. Its implementation principles and technical effects are similar to those described above and will not be repeated here.
第四方面,本發明提供一種圖像處理設備,包括:生成模組,用於生成當前待編碼圖像區塊的重建訊號,並根據重建訊號重建當前待編碼圖像區塊,以得到當前重建圖像區塊;選擇模組,用於若當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則從至少兩個候選濾波器中選擇用於對所述當前重建圖像區塊進行升取樣處理的第一濾波器;處理模組,用於通過第一濾波器對所述當前重建圖像區塊進行升取樣處理。In a fourth aspect, the present invention provides an image processing device, including: a generating module for generating a reconstruction signal of an image block to be encoded, and reconstructing the image block to be encoded according to the reconstruction signal to obtain the current reconstruction Image block; a selection module for selecting the at least two candidate filters for up-sampling the current reconstructed image block if the current reconstruction image block is down-sampled A first filter for sampling processing; a processing module for upsampling the current reconstructed image block through the first filter.
第五方面,本發明提供一種圖像處理設備,包括:解析模組,用於解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和當前待重建圖像區塊的編碼方式;生成模組,用於根據當前待重建圖像區塊的編碼資訊生成當前待重建圖像區塊的重建訊號,並根據重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;選擇模組,用於若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則根據從位元串流獲取的第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,其中,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;處理模組,用於通過第一濾波器對當前重建圖像區塊進行升取樣處理。According to a fifth aspect, the present invention provides an image processing device, including: a parsing module for parsing a bit stream to obtain coding information of a current image block to be reconstructed and a coding method of a current image block to be reconstructed ; Generating module, used to generate the reconstruction signal of the current image block to be reconstructed according to the coding information of the current image block to be reconstructed, and to reconstruct the current image block to be reconstructed according to the reconstruction signal to obtain the current reconstruction image block ; The selection module is used to select upsampling of the currently reconstructed image block according to the first indication information obtained from the bit stream if the current encoding method of the image block to be reconstructed is a down-sampling encoding method; The first filter processed, wherein the first indication information is used to indicate how to select the filter used for upsampling the current reconstructed image block from at least two candidate filters; the processing module is used to pass the A filter performs upsampling on the current reconstructed image block.
第六方面,本發明提供一種圖像處理設備,包括:解析模組,用於解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和當前待重建圖像區塊的編碼方式;生成模組,用於根據編碼資訊生成當前待重建圖像區塊的重建訊號,並根據重建訊號重建所述當前待重建圖像區塊,得到當前重建圖像區塊;解析模組,還用於若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則從位元串流獲取第一濾波器的標識資訊;處理模組,用於通過標識資訊標識的第一濾波器對所述當前重建圖像區塊進行升取樣處理。According to a sixth aspect, the present invention provides an image processing device including: a parsing module for parsing a bit stream to obtain encoding information of a current image block to be reconstructed and encoding method of a current image block to be reconstructed ; Generating module, used to generate the reconstruction signal of the current image block to be reconstructed according to the encoded information, and reconstructing the current image block to be reconstructed according to the reconstruction signal, to obtain the current reconstruction image block; the analysis module, also It is used to obtain the identification information of the first filter from the bit stream if the current encoding method of the image block to be reconstructed is the down-sampling encoding method; the processing module is used to identify the first filter pair identified by the identification information The current reconstructed image block is up-sampled.
第七方面,本發明提供一種圖像處理系統,包括:如第四方面以及第四方面的可選方式所述之圖像處理設備,以及如第五方面以及第五方面的可選方式所述之圖像處理設備。According to a seventh aspect, the present invention provides an image processing system, including: the image processing device according to the fourth aspect and the optional mode of the fourth aspect, and according to the fifth aspect and the optional mode of the fifth aspect Image processing equipment.
第八方面,本發明提供一種圖像處理系統,包括:如第四方面以及第四方面的可選方式所述之圖像處理設備,以及如第六方面以及第六方面的可選方式所述之圖像處理設備。According to an eighth aspect, the present invention provides an image processing system, including: the image processing device as described in the fourth aspect and the optional mode of the fourth aspect, and as described in the sixth aspect and the optional mode of the sixth aspect Image processing equipment.
第九方面,本發明提供一種圖像處理設備,該設備包括經配置以進行以下操作的編碼器:In a ninth aspect, the present invention provides an image processing device including an encoder configured to perform the following operations:
生成當前待編碼圖像區塊的重建訊號,並根據重建訊號重建所述當前待編碼圖像區塊,以得到當前重建圖像區塊;若當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,並通過第一濾波器對當前重建圖像區塊進行升取樣處理。Generate a reconstruction signal of the current image block to be encoded, and reconstruct the current image block to be encoded according to the reconstruction signal to obtain the current reconstruction image block; if the encoding method of the current reconstruction image block is down-sampling encoding In a manner, a first filter for upsampling the current reconstructed image block is selected from at least two candidate filters, and upsampling the current reconstructed image block through the first filter.
第十方面,本發明提供一種圖像處理設備,該設備包括經配置以進行以下操作的解碼器:In a tenth aspect, the present invention provides an image processing device including a decoder configured to perform the following operations:
解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和所述當前待重建圖像區塊的編碼方式;根據當前待重建圖像區塊的編碼資訊生成當前待重建圖像區塊的重建訊號;根據重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則根據從位元串流獲取的第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,其中,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;通過第一濾波器對當前重建圖像區塊進行升取樣處理。Parse the bit stream to obtain the encoding information of the current image block to be reconstructed and the encoding method of the current image block to be reconstructed; generate the current image area to be reconstructed according to the encoding information of the current image block to be reconstructed The reconstruction signal of the block; reconstruct the current image block to be reconstructed according to the reconstruction signal to obtain the current reconstruction image block; if the encoding method of the current image block to be reconstructed is a down-sampling encoding method, it is obtained from the bit stream Of the first indication information selects the first filter for upsampling the current reconstructed image block, wherein the first indication information is used to indicate how to select the current reconstructed image block from at least two candidate filters The filter used in the upsampling process; upsampling the current reconstructed image block through the first filter.
第十一方面,本發明提供一種圖像處理設備,該設備包括經配置以進行以下操作的解碼器:In an eleventh aspect, the present invention provides an image processing device including a decoder configured to perform the following operations:
解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和所述當前待重建圖像區塊的編碼方式;根據編碼資訊生成當前待重建圖像區塊的重建訊號,並根據重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則從位元串流獲取第一濾波器的標識資訊,並通過標識資訊標識的第一濾波器對當前重建圖像區塊進行升取樣處理。Parse the bit stream to obtain the encoding information of the current image block to be reconstructed and the encoding method of the current image block to be reconstructed; generate the reconstruction signal of the current image block to be reconstructed according to the encoding information, and according to the reconstruction The signal reconstructs the current image block to be reconstructed to obtain the current reconstructed image block; if the current image block to be reconstructed is down-sampled, the identification information of the first filter is obtained from the bit stream, And upsampling the current reconstructed image block through the first filter identified by the identification information.
第十二方面,本發明提供一種電腦儲存媒介,用於儲存為上述第四方面或者第九方面涉及的圖像處理設備所用的電腦軟體指令,其包含用於執行上述第四方面或者第九方面所設計的程式。In a twelfth aspect, the present invention provides a computer storage medium for storing computer software instructions used in the image processing apparatus according to the fourth aspect or the ninth aspect, which includes instructions for performing the fourth aspect or the ninth aspect The designed program.
第十三方面,本發明提供一種電腦程式產品,其包含指令,當所述電腦程式被電腦所執行時,該指令使得電腦執行第四方面或者第九方面中圖像處理設備所執行的功能。In a thirteenth aspect, the present invention provides a computer program product including instructions that when executed by a computer, the instructions cause the computer to perform the functions performed by the image processing apparatus of the fourth aspect or ninth aspect.
第十四方面,本發明提供一種電腦儲存媒介,用於儲存為上述第五方面或者第十方面涉及的圖像處理設備所用的電腦軟體指令,其包含用於執行上述第五方面或者第十方面所設計的程式。According to a fourteenth aspect, the present invention provides a computer storage medium for storing computer software instructions for the image processing apparatus according to the fifth aspect or the tenth aspect, which includes instructions for performing the fifth aspect or the tenth aspect The designed program.
第十五方面,本發明提供一種電腦程式產品,其包含指令,當所述電腦程式被電腦所執行時,該指令使得電腦執行第五方面或者第十方面中圖像處理設備所執行的功能。According to a fifteenth aspect, the present invention provides a computer program product including instructions that, when the computer program is executed by a computer, the instructions cause the computer to perform the functions performed by the image processing apparatus of the fifth aspect or the tenth aspect.
第十六方面,本發明提供一種電腦儲存媒介,用於儲存為上述第六方面或者第十一方面涉及的圖像處理設備所用的電腦軟體指令,其包含用於執行上述第五方面或者第十方面所設計的程式。In a sixteenth aspect, the present invention provides a computer storage medium for storing computer software instructions used in the image processing apparatus according to the sixth aspect or the eleventh aspect, which includes instructions for performing the fifth aspect or the tenth aspect The design of the program.
第十七方面,本發明提供一種電腦程式產品,其包含指令,當所述電腦程式被電腦所執行時,該指令使得電腦執行第六方面或者第十一方面中圖像處理設備所執行的功能。According to a seventeenth aspect, the present invention provides a computer program product including instructions that when executed by a computer, the instructions cause the computer to perform the functions performed by the image processing apparatus of the sixth aspect or the eleventh aspect .
本發明提供一種圖像處理方法、設備及系統,相對於現有技術中對整幅圖像中的重建圖像區塊都採用相同的濾波器,本發明針對每個重建圖像區塊都要選擇對應的濾波器,即本發明採用具有針對性的選擇方式來選擇濾波器,通過選擇的濾波器對重建圖像區塊進行升取樣處理,從而可以得到顯示效果更佳的重建圖像區塊。The present invention provides an image processing method, device and system. Compared with the prior art, which uses the same filter for the reconstructed image blocks in the entire image, the present invention must select for each reconstructed image block The corresponding filter, that is, the present invention adopts a targeted selection method to select the filter, and upsamples the reconstructed image block through the selected filter, so that a reconstructed image block with better display effect can be obtained.
為讓本發明的上述特徵和優點能更明顯易懂,下文特舉實施例,並配合所附圖式作詳細說明如下。In order to make the above-mentioned features and advantages of the present invention more obvious and understandable, the embodiments are specifically described below and described in detail in conjunction with the accompanying drawings.
以下,對本發明中的部分專業用語進行解釋說明,以便於本領域技術人員理解。Hereinafter, some technical terms in the present invention will be explained to facilitate understanding by those skilled in the art.
數位影像是以數位形式記錄的影像。數位影像由多序框數位圖像組成。通常原始影像包括的資料量較大,不適合儲存和傳輸,需要利用高效的影像壓縮編碼技術來壓縮原始資料。Digital images are images recorded in digital form. Digital images are composed of multi-order frame digital images. Generally, the original image contains a large amount of data, which is not suitable for storage and transmission. It is necessary to use efficient image compression coding technology to compress the original data.
影像壓縮技術是通過消除影像冗餘來達到壓縮的目的。影像冗餘主要包括如下幾項:空間冗余,時間冗餘,視覺冗餘和資訊亂度冗餘。Image compression technology achieves the purpose of compression by eliminating image redundancy. Image redundancy mainly includes the following items: spatial redundancy, temporal redundancy, visual redundancy and information redundancy.
空間冗餘:它是靜態圖像存在的最主要的資料冗餘。它是指在一幅圖像中,相鄰的畫素的振幅都比較相近,這種空間連貫性被稱為空間相關或空間冗餘。空間冗餘主要是通過序框內預測方法來消除,序框內預測方法是指利用影像空間域的相關性,使用參考重建圖像區塊的畫素預測當前重建圖像區塊的畫素,以達到去除影像空間冗餘的目的。Spatial redundancy: It is the most important data redundancy in the presence of static images. It means that in an image, the amplitudes of adjacent pixels are relatively similar. This spatial coherence is called spatial correlation or spatial redundancy. Spatial redundancy is mainly eliminated by the intra-sequence prediction method. The intra-sequence prediction method refers to using the correlation of the image spatial domain to predict the pixels of the currently reconstructed image block using the pixels of the reference reconstructed image block. In order to achieve the purpose of removing the spatial redundancy of the image.
時間冗餘:它是影像序列中經常包含的冗餘,由於影像的相鄰圖像往往包含相同或類似的背景和運動物體,只是運動物體所在的空間位置略有不同,這種相鄰圖像間的資料的高度相關性就稱為時間冗餘。時間冗餘主要是通過序框間預測技術來消除,序框間預測技術是指利用時間上相鄰圖像的畫素來預測當前畫素。Temporal redundancy: It is often included in the image sequence. Because the adjacent images of the image often contain the same or similar background and moving objects, but the spatial location of the moving objects is slightly different, this adjacent image The high correlation between data is called time redundancy. Temporal redundancy is mainly eliminated by inter-sequence prediction technology. Inter-sequence prediction technology refers to the use of pixels of temporally adjacent images to predict the current pixel.
視覺冗餘:人眼視覺系統對圖像細節的變化不敏感,這些細微變化資訊即使丟失,人眼也感受不到。而在記錄原始影像資料時,通常假定視覺系統是對各種內容的敏感度是一致的,這樣就產生了比理想編碼更多的資料,稱為視覺冗餘。視覺冗餘主要是通過變換,量化技術來消除,該變換技術是指將圖像訊號變換到頻率域進行處理,根據不同頻率訊號對視覺品質的貢獻大小進行資料表達和位元再分配,這樣可以糾正空間域上均勻取樣的不合理表達。同時在位元再分配過程中融合考慮去除視覺冗餘的需要,通過量化操作,省略過分精細的高頻分量表達,實現有效壓縮。Visual redundancy: The human visual system is insensitive to changes in image details. Even if these subtle changes are lost, the human eye will not feel it. When recording original image data, it is usually assumed that the visual system is consistent in sensitivity to various contents, which produces more data than ideal coding, called visual redundancy. Visual redundancy is mainly eliminated by transformation and quantization techniques. This transformation technique refers to transforming the image signal into the frequency domain for processing, and performing data expression and bit redistribution according to the contribution of different frequency signals to visual quality. Correct the unreasonable expression of uniform sampling in the spatial domain. At the same time, in the process of bit redistribution, fusion needs to consider the need to remove visual redundancy. Through the quantization operation, the excessively high-frequency component expression is omitted, and effective compression is achieved.
訊息亂度冗餘:由資訊理論可知,為表示圖像資料的一個畫素,只要按其資訊亂度的大小分配相應位元數即可,而對於圖像資料的每個畫素,在圖像獲取時很難得到它的資訊亂度,因此一般是對每個畫素採用相同的位元數來表示,這樣必然存在冗餘。資訊亂度冗餘主要是通過亂度編碼技術來消除亂度編碼技術是通過統計係數的資訊亂度分佈,為具有不同資訊亂度的資料分配不同的位元數。Information disorder redundancy: According to information theory, for a pixel representing image data, it is sufficient to allocate the corresponding number of bits according to the size of its information disorder. For each pixel of image data, For example, it is difficult to get its information disorder when acquiring, so it is generally represented by the same number of bits for each pixel, so there must be redundancy. Information scrambling redundancy is mainly eliminated through scrambling coding technology. The coding technology is based on the distribution of information scrambling of statistical coefficients to allocate different bits for data with different information scrambling.
當前主流的影像壓縮編碼架構是混合編碼架構,針對上述冗餘,採取不同的技術來消除冗餘,並將這些技術結合在一起,形成了影像編碼的混合架構。如圖3所示,編碼端接收到影像之後,對於構成影像的每序框圖像,將該圖像劃分成待編碼圖像區塊。對於當前待編碼圖像區塊,首先通過參考重建圖像區塊對當前待編碼圖像區塊進行預測,得到當前待編碼圖像區塊的預測訊號;用當前待編碼圖像區塊的原始訊號減去預測訊號,得到殘差訊號。經過預測後,殘差訊號的振幅遠小於原始訊號。將殘差訊號進行變換和量化操作。經過變換量化後,得到變換量化係數,再通過亂度編碼技術編碼量化係數以及編碼中的其他指示資訊,得到位元串流。進一步地,編碼端還需要重建當前待編碼圖像區塊,以實現對後續待編碼圖像區塊的編碼提供參考畫素。具體地,在得到當前待編碼圖像區塊的變換量化係數之後,編碼端需要對當前待編碼圖像區塊的變換量化係數進行反量化和反變換,得到重建的殘差訊號,將重建的殘差訊號與當前待編碼圖像區塊對應的預測訊號相加,得到當前待編碼圖像區塊的重建訊號,根據該重建訊號得到重建圖像區塊。The current mainstream image compression coding architecture is a hybrid coding architecture. For the above-mentioned redundancy, different technologies are adopted to eliminate the redundancy, and these technologies are combined to form a hybrid architecture for image coding. As shown in FIG. 3, after receiving the video, the encoding end divides the image into image blocks to be coded for each sequence frame image constituting the video. For the current image block to be encoded, first predict the current image block to be encoded by referring to the reconstructed image block to obtain the prediction signal of the current image block to be encoded; using the original image block to be encoded The signal is subtracted from the predicted signal to obtain the residual signal. After prediction, the amplitude of the residual signal is much smaller than the original signal. Transform and quantize the residual signal. After transform and quantization, transform and quantization coefficients are obtained, and then the quantization coefficients and other indication information in the encoding are encoded by the random coding technology to obtain a bit stream. Further, the encoding end also needs to reconstruct the current image block to be encoded, so as to provide reference pixels for encoding the subsequent image block to be encoded. Specifically, after obtaining the transform and quantization coefficients of the current image block to be encoded, the encoding end needs to inversely quantize and inverse transform the transform and quantization coefficients of the current image block to be encoded, to obtain a reconstructed residual signal, and then reconstruct the reconstructed residual signal. The residual signal is added to the prediction signal corresponding to the current image block to be encoded to obtain the reconstruction signal of the current image block to be encoded, and the reconstructed image block is obtained according to the reconstruction signal.
如圖4所示,解碼端獲取到位元串流之後,首先對位元串流進行亂度解碼,得到當前待重建圖像區塊的變換量化係數,然後對變換量化係數進行反量化和反變換,得到當前待重建圖像區塊的重建的殘差訊號。通過參考重建圖像區塊對當前待重建圖像區塊進行預測,得到當前待重建圖像區塊的預測訊號,然後將預測訊號和上述重建的殘差訊號相加,得到當前待重建圖像區塊的重建訊號,然後根據該重建訊號得到當前待重建圖像區塊對應的當前重建圖像區塊。As shown in FIG. 4, after acquiring the bit stream, the decoding end first decodes the bit stream randomly to obtain the transform and quantization coefficients of the image block to be reconstructed, and then inversely quantizes and inverse transforms the transform and quantization coefficients To obtain the reconstructed residual signal of the current image block to be reconstructed. Predict the current image block to be reconstructed by referring to the reconstructed image block to obtain the predicted signal of the current image block to be reconstructed, and then add the predicted signal and the reconstructed residual signal to obtain the current image to be reconstructed The reconstruction signal of the block, and then obtain the current reconstruction image block corresponding to the current image block to be reconstructed according to the reconstruction signal.
為了降低編解碼複雜度,編碼端對每序框圖像先進行縮減取樣處理,如圖5所示,編碼端對整幅圖像採用縮減取樣處理,然後對經過縮減取樣處理後的圖像中的每個待編碼圖像區塊進行編碼,得到位元串流。其中,每個待編碼圖像區塊對應的重建圖像區塊的解析度為縮減取樣解析度。相應的,解碼端解析位元串流,每個待重建圖像區塊的解析度均為縮減取樣解析度,得到對應的重建圖像區塊的解析度也為縮減取樣解析度,解碼端需要對重建圖像區塊採用升取樣處理,以得到原始解析度的重建圖像區塊。In order to reduce the complexity of encoding and decoding, the encoding end performs downsampling on each sequence frame image. As shown in Figure 5, the encoding end uses downsampling on the entire image, and then on the image after downsampling Each image block to be encoded is encoded to obtain a bit stream. The resolution of the reconstructed image block corresponding to each image block to be encoded is the down-sampling resolution. Correspondingly, the decoding end parses the bit stream. The resolution of each image block to be reconstructed is the down-sampling resolution. The resolution of the corresponding reconstructed image block is also the down-sampling resolution. The decoder needs to Upsampling is applied to the reconstructed image block to obtain the reconstructed image block with original resolution.
其中,在編碼端和解碼端均涉及通過參考重建圖像區塊對當前重建圖像區塊(當前待編碼圖像區塊或者當前待重建圖像區塊)進行預測,得到當前重建圖像區塊的預測訊號。在本發明中,對當前重建圖像區塊的預測模式(主要是序框內預測方法)可以採用現有技術,具體如下:Among them, both the encoding end and the decoding end involve predicting the current reconstructed image block (currently to be encoded image block or current to be reconstructed image block) by referring to the reconstructed image block to obtain the current reconstructed image area The prediction signal of the block. In the present invention, the prediction mode (mainly the intra-sequence prediction method) of the currently reconstructed image block can adopt the existing technology, as follows:
例如:圖6為本發明一實施例提供的正在編碼的圖像的示意圖,如圖6所示,該圖像包括多個圖像區塊,其中,該圖像的編碼順序是:從上到下,從左到右。在圖6中,圖像區塊C、B、D、E和A表示已經完成重建的重建圖像區塊,圖像區塊F為當前待編碼圖像區塊,該圖像中的其他區域為未編碼的圖像區域。For example: FIG. 6 is a schematic diagram of an image being encoded according to an embodiment of the present invention. As shown in FIG. 6, the image includes multiple image blocks, where the encoding order of the image is: from top to Next, from left to right. In FIG. 6, image blocks C, B, D, E, and A represent the reconstructed image blocks that have been reconstructed, image block F is the current image block to be encoded, and other areas in the image It is an uncoded image area.
這裡以H.265標準來說明序框內預測方法的具體過程,H.265支援將當前待編碼圖像區塊劃分成更小的子圖像區塊進行預測操作。子圖像區塊的劃分結構為四元樹結構,即一個圖像區塊可以劃分成四個子圖像區塊,每個子圖像區塊可以繼續劃分成四個子圖像區塊。如圖6所示,假設當前待編碼圖像區塊被劃分成7個子圖像區塊進行預測操作,當前待編碼圖像區塊也可以劃分成更多的子圖像區塊進行預測操作。對於每個子圖像區塊,先進行預測操作,獲得預測訊號,然後根據預測訊號獲得子圖像區塊的殘差訊號,進一步對殘差訊號進行變換、量化和亂度編碼。對於預測操作,每個子圖像區塊可選的序框內預測方法有35種,包括Planar模式、DC模式以及33種角度預測模式。所有預測模式都使用相同的參考畫素範本(由多個參考畫素構成),圖7為本發明一實施例提供的參考畫素範本的示意圖,如圖7所示,P1,1 ,P2,1 ……PN,1 ……P1,N ,P2,N ……PN,N 這些畫素構成待編碼的子圖像區塊,例如:該待編碼的子圖像區塊可以是圖6中的子圖像區塊1。如圖7所示,除了待編碼的子圖像區塊,其他參考畫素R0,0 ,R1,0 ……R2N+1,0 ……R0,2N 構成參考畫素範本,假設該待編碼的子圖像區塊是圖6中的子圖像區塊1,這種情況下,這些參考畫素中的一部分畫素為參考重建圖像區塊B的最後一行的畫素,另一部分畫素為參考重建圖像區塊A的最右側一列的畫素。對於其他標準,這些參考畫素中的一部分畫素為參考重建圖像區塊B包括的下方多行的畫素,另一部分畫素為參考重建圖像區塊A包括的右側多列的畫素。即本發明對參考畫素範本不做限制。The H.265 standard is used here to explain the specific process of the intra-frame prediction method. H.265 supports dividing the current image block to be coded into smaller sub-image blocks for prediction operations. The division structure of the sub-image blocks is a quaternary tree structure, that is, one image block can be divided into four sub-image blocks, and each sub-image block can be further divided into four sub-image blocks. As shown in FIG. 6, assuming that the current image block to be encoded is divided into 7 sub-image blocks for prediction operation, the current image block to be encoded may also be divided into more sub-image blocks for prediction operation. For each sub-image block, first perform a prediction operation to obtain a prediction signal, and then obtain a residual signal of the sub-image block according to the prediction signal, and further transform, quantize, and encode the residual signal. For the prediction operation, there are 35 selectable in-frame prediction methods for each sub-image block, including Planar mode, DC mode, and 33 angle prediction modes. All prediction modes use the same reference pixel template (composed of multiple reference pixels). FIG. 7 is a schematic diagram of a reference pixel template provided by an embodiment of the present invention. As shown in FIG. 7, P 1,1 , P 2,1 ...... P N, 1 ...... P 1, N , P 2, N ...... P N, N These pixels constitute the sub-image block to be encoded, for example: the sub-image block to be encoded It may be the sub-image block 1 in FIG. 6. As shown in FIG. 7, in addition to the sub-image block to be encoded, other reference pixels R 0,0 , R 1,0 ... R 2N + 1,0 ... R 0,2N constitute the reference pixel template, assuming The sub-picture block to be coded is sub-picture block 1 in FIG. 6, in this case, some of the reference pixels are the pixels in the last row of the reference reconstruction image block B The other part of the pixels is the pixel in the rightmost column of the reference reconstruction image block A. For other standards, some of these reference pixels are the pixels in the lower rows included in the reference reconstruction image block B, and the other pixels are the pixels in the right column included in the reference reconstruction image block A . That is, the present invention does not limit the reference pixel template.
Planar模式Planar mode
Planar模式適用於畫素值緩慢變化的區域,圖8A和圖8B為本發明一實施例提供的Planar模式的示意圖,如圖8所示,使用水平和垂直方向的兩個線性濾波器,分別得到兩個預測值和,並將和的平均值作為畫素的預測訊號。Planar mode is suitable for areas with slowly changing pixel values. Figures 8A and 8B are schematic diagrams of the Planar mode provided by an embodiment of the present invention. As shown in Figure 8, two linear filters in the horizontal and vertical directions are used to obtain Two predicted values with And with The average value of pixels Forecast signal.
DC模式DC mode
DC模式適用於大面積平坦區域,當前待編碼的子圖像區塊的預測訊號可由其左側和上方的參考畫素的平均值得到,如圖7所示,待編碼的子圖像區塊中各個畫素的預測訊號可以通過的平均值得到。The DC mode is suitable for large flat areas. The prediction signal of the current sub-picture block to be encoded can be obtained from the average value of the reference pixels on the left and above, as shown in FIG. 7, in the sub-picture block to be encoded The prediction signal of each pixel can pass The average value is obtained.
角度模式Angle mode
H.265/HEVC規定了33種角度預測模式,以更好地適應影像內容中不同方向的紋理。圖9為本發明一實施例提供的33種角度預測模式的具體方向示意圖,如圖9所示,33種角度預測模式分為水平類模式(2~17)和垂直類模式(18~34)。其中V0(模式26)和H0(模式10)分別表示垂直和水平方向,其餘角度預測模式的預測方向都可以看作是在垂直或水平方向上做一個角度偏移。這裡以垂直方向V0(26)為例說明角度預測過程,垂直方向預測是使用當前待編碼的子圖像區塊上方相鄰的一行參考畫素來預測當前待編碼的子圖像區塊,當前待編碼的子圖像區塊內每個畫素的預測訊號等於該畫素所在列對應的參考畫素的畫素值,即。對於其他角度預測模式,會與水平或垂直方向有一個角度偏移,根據這個角度偏移可以計算參考畫素的位置。該參考畫素的位置可能是兩個相鄰參考畫素之間的位置,若是這種情況,則需要在兩個參考畫素之間根據計算出來的位置插值得到一個參考畫素。通過得到的參考畫素生成預測訊號。H.265 / HEVC stipulates 33 angle prediction modes to better adapt to the texture in different directions in the video content. FIG. 9 is a schematic diagram of specific directions of 33 angle prediction modes provided by an embodiment of the present invention. As shown in FIG. 9, the 33 angle prediction modes are divided into horizontal mode (2-17) and vertical mode (18-34) . Among them, V0 (mode 26) and H0 (mode 10) represent the vertical and horizontal directions, respectively, and the prediction directions of the other angle prediction modes can be regarded as an angle offset in the vertical or horizontal direction. Here, the vertical direction V0 (26) is used as an example to explain the angle prediction process. The vertical direction prediction uses the adjacent row of reference pixels above the sub-picture block to be encoded to predict the sub-picture block to be encoded. The prediction signal of each pixel in the encoded sub-image block is equal to the pixel value of the reference pixel corresponding to the column of the pixel, ie . For other angle prediction modes, there will be an angle offset from the horizontal or vertical direction, and the position of the reference pixel can be calculated according to this angle offset. The position of the reference pixel may be a position between two adjacent reference pixels. If this is the case, a reference pixel needs to be interpolated between the two reference pixels according to the calculated position. The predicted signal is generated by the obtained reference pixels.
需要說明的是,上述序框內預測方法同樣適用於解碼端,本發明在此不再贅述。It should be noted that the foregoing in-frame prediction method is also applicable to the decoding end, and the present invention will not repeat them here.
本發明還涉及到圖像縮減取樣處理和圖像升取樣處理。The invention also relates to image downsampling processing and image upsampling processing.
其中,圖像縮減取樣處理涉及三個方面的資訊:1、縮減取樣比例;2、縮減取樣位置;3、縮減取樣所使用的濾波器。Among them, the image downsampling process involves three aspects of information: 1. Downsampling ratio; 2. Downsampling position; 3. Filter used for downsampling.
縮減取樣比例是指原圖像與縮減取樣後的圖像的比例,可以分別在水平方向和垂直方向來描述。例如可以對圖像訊號進行水平方向2:1縮減取樣,垂直方向4:1縮減取樣;或者水平方向不縮減取樣,垂直方向2:1縮減取樣;或者水平和垂直方向都進行2:1縮減取樣等。The down-sampling ratio refers to the ratio of the original image to the down-sampled image, which can be described in the horizontal direction and the vertical direction, respectively. For example, the image signal can be downsampled horizontally 2: 1, downsampled vertically 4: 1; or not downsampled horizontally, downsampled vertically 2: 1; or both horizontally and vertically downsampled 2: 1 Wait.
縮減取樣位置是指縮減取樣點與原取樣點的位置關係,例如,縮減取樣點的位置可以與部分原取樣點位置一樣,或者縮減取樣點落在幾個原取樣點之間。The down-sampling position refers to the position relationship between the down-sampling point and the original sampling point. For example, the position of the down-sampling point may be the same as the position of some original sampling points, or the down-sampling point falls between several original sampling points.
縮減取樣濾波器可以是3-lobe Lanczos濾波器,Bilinear濾波器,Bicubic,Gauss濾波器等。The down-sampling filter can be a 3-lobe Lanczos filter, Bilinear filter, Bicubic, Gauss filter, etc.
下面以解析度為16*16的圖像區塊(實際的圖像會比這大很多,例如1920*1080)為例來說明縮減取樣過程。圖10為本發明一實施例提供的圖像縮減取樣示意圖,假設水平方向和垂直方向的取樣比例均為2:1,水平方向上,縮減取樣點的位置落在兩個原取樣點的左側原取樣點位置,垂直方向上,縮減取樣點落在兩個原取樣點的上方原取樣點位置。如圖10所示,框出來的圓圈表示縮減取樣點的位置,縮減取樣的濾波器如下:The following uses an image block with a resolution of 16 * 16 (the actual image will be much larger than this, such as 1920 * 1080) as an example to illustrate the downsampling process. 10 is a schematic diagram of image downsampling provided by an embodiment of the present invention. Assuming that the sampling ratio in the horizontal direction and the vertical direction is 2: 1, in the horizontal direction, the position of the downsampling point falls to the left of the two original sampling points. The sampling point position, in the vertical direction, the reduced sampling point falls above the two original sampling points. As shown in Figure 10, the circled frame indicates the position of the downsampling point. The downsampling filter is as follows:
該濾波器是簡單的低通濾波器,該低通濾波器可以被看作是一個二維濾波器,也可以被看作是兩個一維濾波器。如果當作一個二維濾波器,可以在一次濾波操作中同時完成水平和垂直方向上的縮減取樣。如圖10所示,對縮減取樣點A進行縮減取樣時,用到了鄰近的8個原取樣點(用三角形框起的圓圈),根據上面濾波器計算出該縮減取樣點A的畫素值。如果當作是兩個一維濾波器,需要先完成水平或者垂直方向的縮減取樣,再對已完成水平或者垂直方向縮減取樣的結果進行垂直或水平方向縮減取樣。如圖10所示,對縮減取樣點A進行縮減取樣時,先利用縮減取樣點A左右各一個原取樣點進行水平縮減取樣,然後對縮減取樣後的結果利用縮減取樣點A上下各一個原取樣點進行垂直方向上的縮減取樣,根據上述濾波器計算出縮減取樣點A的畫素值。採用相同的方法,對整個16*16的圖像區塊進行縮減取樣處理,最終縮減取樣的結果如圖10所示,各個縮減取樣點的位置如框出來的圓圈所示,縮減取樣點的畫素值為通過濾波器操作後的數值。如圖10所示,縮減取樣後的圖像區塊的解析度為8*8。The filter is a simple low-pass filter, which can be regarded as a two-dimensional filter or two one-dimensional filters. If it is used as a two-dimensional filter, the downsampling in the horizontal and vertical directions can be completed in one filtering operation. As shown in FIG. 10, when downsampling the downsampling point A, 8 adjacent original sampling points (circles surrounded by triangles) are used, and the pixel value of the downsampling point A is calculated according to the above filter. If it is regarded as two one-dimensional filters, it is necessary to complete the downsampling in the horizontal or vertical direction first, and then perform the downsampling in the vertical or horizontal direction on the results of the completed downsampling in the horizontal or vertical direction. As shown in FIG. 10, when performing downsampling on the downsampling point A, first use one of the original sampling points on the left and right of the downsampling point A to perform horizontal downsampling, and then use the downsampling point A on each of the original sampling points The points are down-sampled in the vertical direction, and the pixel value of the down-sampled point A is calculated according to the above filter. Using the same method, the entire 16 * 16 image block is down-sampled. The final result of the down-sampling is shown in Figure 10. The position of each down-sampling point is shown by the circled frame. The prime value is the value after the filter operation. As shown in FIG. 10, the resolution of the image block after downsampling is 8 * 8.
通常編碼端或者解碼端需要對經過縮減取樣後的圖像進行升取樣處理,目的是為了得到原解析度的圖像。升取樣處理涉及三個方面的資訊:1、升取樣比例;2、升取樣位置;3升取樣所使用的濾波器。Usually the encoder or decoder needs to up-sample the down-sampled image in order to get the original resolution image. The upsampling process involves three aspects of information: 1. Upsampling ratio; 2. Upsampling position; 3 Filter used for upsampling.
升取樣比例是指升取樣前的圖像與升取樣後的圖像的比例,可以分別在水平方向和垂直方向來描述。例如可以對升取樣才的圖像訊號進行水平方向1:2升取樣,垂直方向1:4升取樣;或者水平方向不升取樣,垂直方向1:2升取樣;或者水平和垂直方向都進行1:2升取樣等。The upsampling ratio refers to the ratio of the image before upsampling to the image after upsampling, which can be described in the horizontal direction and the vertical direction, respectively. For example, the up-sampling image signal can be up-sampled by 1: 2 in the horizontal direction and up-sampled by 1: 4 in the vertical direction; or not up-sampled in the horizontal direction, up-sampled by 1: 2 in the vertical direction; or both horizontal and vertical : 2 liter sampling etc.
升取樣位置是指升取樣後的取樣點與升取樣前的取樣點的位置關係,例如,圖11為本發明一實施例提供的圖像升取樣示意圖,如圖11所示,在第一行,採用水平1:2升取樣比例,升取樣後的取樣點的位置可以在升取樣前的取樣點的右側,其中×表示升取樣後的取樣點位置,圓圈表示升取樣前的取樣點位置。在第二行,採用水平1:2升取樣比例,升取樣後的取樣點的位置可以在升取樣前的取樣點的左側,其中×表示升取樣後的取樣點位置,圓圈表示升取樣前的取樣點位置。需要說明的是,升取樣後的取樣點的位置應與上述縮減取樣點的位置選取相對應,例如:在進行縮減取樣時,選擇縮減取樣點的位置為它的左側原取樣點的位置,那麼在進行升取樣時,選擇升取樣後的取樣點的位置為它的右側升取樣前的取樣點(縮減取樣點)的位置。The upsampling position refers to the positional relationship between the sampling point after upsampling and the sampling point before upsampling. For example, FIG. 11 is a schematic diagram of image upsampling provided by an embodiment of the present invention. As shown in FIG. 11, the first line The horizontal 1: 2 upsampling ratio is adopted. The position of the sampling point after upsampling can be on the right side of the sampling point before upsampling, where × indicates the position of the sampling point after upsampling and the circle indicates the position of the sampling point before upsampling. In the second line, the horizontal 1: 2 upsampling ratio is adopted. The position of the sampling point after upsampling can be on the left side of the sampling point before upsampling, where × indicates the position of the sampling point after upsampling, and the circle indicates the position before sampling. Sampling point location. It should be noted that the position of the sampling point after upsampling should correspond to the above selection of the downsampling point, for example: when performing downsampling, select the position of the downsampling point as the position of the original sampling point on its left, then When upsampling, select the position of the sampling point after upsampling as the position of the sampling point (downsampling point) before upsampling on the right side.
升取樣濾波器可以是DCTIF濾波器,雙線性插值濾波器,sinc濾波器等。下面以解析度為8*8的圖像區塊(即上述縮減取樣後的圖像區塊)為例來說明升取樣過程。假設水平方向和垂直方向的升取樣比例均為1:2,在水平方向上,升取樣後的取樣點的位置為右側升取樣前的取樣點的位置,在垂直方向上,升取樣後的取樣點的位置為下方升取樣前的取樣點的位置,這裡以水平方向和垂直方向分別升取樣為例,以及以DCTIF濾波器為例來說明升取樣處理過程。DCTIF濾波器為(-1, 4,-11, 40, 40, -11, 4, -1 ),假設當前要進行水平方向升取樣,在圖11中,假設需要***B3取樣點,則採用如下公式確定B3的畫素值:The upsampling filter can be a DCTIF filter, a bilinear interpolation filter, a sinc filter, etc. The following uses the image block with a resolution of 8 * 8 (that is, the image block after downsampling) as an example to explain the upsampling process. Assuming that the horizontal and vertical upsampling ratios are 1: 2, in the horizontal direction, the position of the sampling point after upsampling is the position of the sampling point on the right before upsampling, and in the vertical direction, the sampling after upsampling The position of the point is the position of the sampling point before the upsampling below. Here, the upsampling in the horizontal and vertical directions is taken as an example, and the DCTIF filter is used as an example to explain the upsampling process. The DCTIF filter is (-1, 4, -11, 40, 40, -11, 4, -1), assuming that horizontal upsampling is currently being performed. In Figure 11, assuming that the B3 sampling point needs to be inserted, the following is used. The formula determines the pixel value of B3:
對於其他位置的插值取樣點,例如B7,需要用到B7右邊的四個畫素,這些畫素目前不可得,實際中一般會將A7重複4次,以用於計算B7的畫素值。在垂直方向上的升取樣與在水平方向上的升取樣類似,在此不再贅述。也可以先進行垂直方向上的升取樣,再進行水平方向上的升取樣。圖12為本發明一實施例提供的圖像升取樣示意圖,如圖12所示,×表示升取樣後的取樣點,圓圈表示升取樣前的取樣點。For interpolation sampling points at other positions, such as B7, the four pixels on the right of B7 are needed. These pixels are currently unavailable. In practice, A7 is generally repeated 4 times to calculate the pixel value of B7. The upsampling in the vertical direction is similar to the upsampling in the horizontal direction, and will not be repeated here. Upsampling in the vertical direction can also be performed first, followed by upsampling in the horizontal direction. FIG. 12 is a schematic diagram of image upsampling provided by an embodiment of the present invention. As shown in FIG. 12, X indicates a sampling point after upsampling, and a circle indicates a sampling point before upsampling.
升取樣濾波器還可以是6抽頭的維納Wiener濾波器,通過濾波器對圖像進行升取樣處理的過程如下:圖13為本發明一實施例提供的升取樣圖像示意圖,如圖13所示,圖中沒有標號的方框表示整數畫素,aa,bb,cc,dd,ee,ff,gg,hh和b,h,s,m,j均是1/2位置畫素,其他都是1/4位置畫素。計算過程是先用(1,-5,20,20,-5,1)的六抽頭濾波器進行1/2位置畫素的插值,然後再通過臨近畫素插值的方法計算1/4位置畫素的插值,從而獲得最終升取樣圖像。畫素有三種類型:水平畫素、垂直畫素和對角線畫素。具體如下:The upsampling filter may also be a 6-tap Wiener Wiener filter. The process of upsampling the image through the filter is as follows: FIG. 13 is a schematic diagram of an upsampling image provided by an embodiment of the present invention, as shown in FIG. 13 As shown, the unmarked boxes in the figure represent integer pixels, aa, bb, cc, dd, ee, ff, gg, hh and b, h, s, m, j are all 1 / 2-position pixels, and others It is a 1/4 position pixel. The calculation process is to use the (1, -5, 20, 20, -5, 1) six-tap filter to interpolate the 1/2 position pixels, and then calculate the 1/4 position pixels by interpolating adjacent pixels Element interpolation to obtain the final up-sampled image. There are three types of pixels: horizontal pixels, vertical pixels and diagonal pixels. details as follows:
水平1/2位置畫素:如b=(E-5F+20G+20H-5I+J),b=Clip1((b+16)>>5);Clip1的作用是限制結果在0~255,右移5相當於除以32,加16是為了結果的四捨五入。Horizontal 1/2 pixel: such as b = (E-5F + 20G + 20H-5I + J), b = Clip1 ((b + 16) >> 5); Clip1 ’s role is to limit the result to 0 ~ 255, Shifting 5 to the right is equivalent to dividing by 32, and adding 16 is for rounding the result.
垂直1/2位置畫素:如h=(A-5C+20G+20M-5R+T),h=Clip1((h+16)>>5);Vertical 1/2 position pixel: such as h = (A-5C + 20G + 20M-5R + T), h = Clip1 ((h + 16) >> 5);
對角1/2位置畫素:如j=(cc-5dd+20h+20m-5ee+ff)=(aa-5bb+20b+20q-5gg+hh),j=Clip1((j+16)>>5);Pixels at the diagonal 1/2 position: such as j = (cc-5dd + 20h + 20m-5ee + ff) = (aa-5bb + 20b + 20q-5gg + hh), j = Clip1 ((j + 16)> > 5);
水平1/4位置畫素:如a=(G+b+1)>>1;i=(h+j+1)>>1;Horizontal 1/4 position pixels: such as a = (G + b + 1) >> 1; i = (h + j + 1) >> 1;
垂直1/4位置畫素:如d=(G+h+1)>>1;f=(b+j+1)>>1;Vertical 1/4 position pixel: if d = (G + h + 1) >> 1; f = (b + j + 1) >> 1;
對角1/4位置畫素:如e=(h+b+1)>>1;g=(b+m+1)>>1;p=(h+s+1)>>1;r=(s+m+1)>>1。Pixels at diagonal 1/4 position: e = (h + b + 1) >> 1; g = (b + m + 1) >> 1; p = (h + s + 1) >> 1; r = (s + m + 1) >> 1.
現有技術中,編碼端或者解碼端對於整幅圖像中每個重建圖像區塊均採用相同的濾波器進行升取樣處理,然而,各個重建圖像區塊的特性可能不同,例如,有些重建圖像區塊可能較為平坦;有些重建圖像區塊可能細節比較多,現有技術勢必導致有些經過升取樣處理之後的重建圖像區塊比較模糊,存在顯示效果不佳的問題。In the prior art, the encoding end or the decoding end uses the same filter for upsampling processing for each reconstructed image block in the entire image, however, the characteristics of each reconstructed image block may be different, for example, some reconstructions The image blocks may be relatively flat; some reconstructed image blocks may have more details. The prior art will inevitably cause some reconstructed image blocks that have undergone upsampling to be blurred and have a problem of poor display effect.
為了解決上述技術問題,本發明提供一種圖像處理方法、設備及系統,本發明可以基於如圖3和圖5的編碼示意圖,如圖3和圖5所示,一幅圖像包括的待編碼圖像區塊的編碼方式可以是如圖3所示的原解析度編碼方式或者是如圖5所示的縮減取樣編碼方式。其中,原解析度編碼方式是指對當前待編碼圖像區塊直接進行編碼操作。縮減取樣編碼方式是指對當前待編碼圖像區塊先進行縮減取樣處理,然後對縮減取樣後的當前待編碼圖像區塊進行編碼操作。一般情況下,紋理圖像區塊採用原解析度編碼方式,平滑圖像區塊採用縮減取樣編碼方式。本發明的主旨思想在於:若當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則為當前重建圖像區塊選擇濾波器,並通過該濾波器對當前重建圖像區塊進行升取樣處理,即針對不同的重建圖像區塊,都要選擇濾波器,以進行升取樣處理。In order to solve the above technical problems, the present invention provides an image processing method, device and system. The present invention can be based on the encoding schematic diagrams shown in FIGS. 3 and 5, as shown in FIGS. 3 and 5, an image includes the to-be-encoded The coding method of the image block may be the original resolution coding method shown in FIG. 3 or the down-sampling coding method shown in FIG. 5. The original resolution encoding method refers to directly encoding the current image block to be encoded. The down-sampling encoding method refers to first performing down-sampling processing on the current image block to be encoded, and then performing encoding operation on the down-sampling current image block to be encoded. Generally, the texture image block adopts the original resolution coding method, and the smooth image block uses the down-sampling coding method. The main idea of the present invention is that if the current reconstruction image block encoding method is a down-sampling encoding method, a filter is selected for the current reconstruction image block, and the current reconstruction image block is up-sampled by the filter Processing, that is, for different reconstructed image blocks, filters must be selected for upsampling.
具體地,圖14為本發明一實施例提供的一種圖像處理方法的流程圖,如圖14所示,該方法包括:Specifically, FIG. 14 is a flowchart of an image processing method according to an embodiment of the present invention. As shown in FIG. 14, the method includes:
步驟S1401:生成當前待編碼圖像區塊的重建訊號,並根據重建訊號重建當前待編碼圖像區塊,以得到當前重建圖像區塊;Step S1401: generate a reconstruction signal of the current image block to be encoded, and reconstruct the current image block to be encoded according to the reconstruction signal to obtain the current reconstruction image block;
步驟S1402:若當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,並通過第一濾波器對當前重建圖像區塊進行升取樣處理。Step S1402: If the current reconstruction image block encoding method is a down-sampling encoding method, select a first filter for upsampling the current reconstruction image block from at least two candidate filters, and pass The first filter performs upsampling on the current reconstructed image block.
具體地,在步驟S1401中:編碼端可以獲取當前圖像的當前待編碼圖像區塊的編碼方式和每個參考重建圖像區塊中的畫素;根據當前待編碼圖像區塊的編碼方式和M個參考重建圖像區塊中的畫素,確定當前待編碼圖像區塊的多個參考畫素;根據多個參考畫素生成當前待編碼圖像區塊的預測訊號;獲取當前待編碼圖像區塊的編碼訊號,其中,當當前待編碼圖像區塊的編碼方式是原解析度編碼方式時,編碼訊號為當前待編碼圖像區塊的原始訊號,當當前待編碼圖像區塊的編碼方式是縮減取樣編碼方式時,編碼訊號為當前待編碼圖像區塊的原始訊號經過縮減取樣處理後得到的訊號;根據預測訊號和編碼訊號生成當前待編碼圖像區塊的殘差訊號;將殘差訊號進行變換和量化操作。經過變換量化後,得到變換量化係數,編碼端對當前待編碼圖像區塊的變換量化係數進行反量化和反變換,得到重建殘差訊號,將重建殘差訊號與當前待編碼圖像區塊對應的預測訊號相加,得到當前待編碼圖像區塊的重建訊號,根據該重建訊號得到當前重建圖像。Specifically, in step S1401: the encoding end may acquire the encoding method of the current image block to be encoded of the current image and the pixels in each reference reconstruction image block; according to the encoding of the current image block to be encoded Method and M reference pixels in the reconstructed image block to determine the multiple reference pixels of the current image block to be encoded; generate the prediction signal of the current image block to be encoded according to the multiple reference pixels; obtain the current The encoding signal of the image block to be encoded, wherein, when the current encoding method of the image block to be encoded is the original resolution encoding method, the encoding signal is the original signal of the current image block to be encoded, when the current image to be encoded When the coding method of the image block is the down-sampling coding method, the coding signal is the signal obtained after the down-sampling process of the original signal of the current image block to be coded; the current image block to be coded is generated according to the prediction signal and the coding signal Residual signal; transform and quantize the residual signal. After transform and quantization, transform and quantization coefficients are obtained, and the encoding end inversely quantizes and inverse transforms the transform and quantization coefficients of the current image block to be encoded to obtain a reconstructed residual signal. The corresponding prediction signals are added to obtain the reconstruction signal of the current image block to be encoded, and the current reconstruction image is obtained according to the reconstruction signal.
其中,當前待重建圖像區塊對應M個參考重建圖像區塊,M為大於或者等於1的正整數。參考重建圖像區塊用於確定待重建圖像區塊的多個參考畫素,其中,多個參考畫素用於生成當前重建圖像區塊的預測訊號。實際上,參考重建圖像區塊具體是哪個重建圖像區塊和解碼端採用的預測模式有關。當採用上述35中預測模式中的任一種時,可以參考如圖7所示的參考畫素範本。Wherein, the current image block to be reconstructed corresponds to M reference image blocks, M is a positive integer greater than or equal to 1. The reference reconstructed image block is used to determine multiple reference pixels of the image block to be reconstructed, wherein the multiple reference pixels are used to generate a prediction signal of the current reconstructed image block. In fact, which specific reconstruction image block is referenced to the reconstruction image block is related to the prediction mode adopted by the decoding end. When any of the above 35 prediction modes is used, the reference pixel template as shown in FIG. 7 can be referred to.
實際上,參考重建圖像區塊具體是哪個重建圖像區塊和解碼端採用的預測模式有關。當採用上述35中預測模式中的任一種時,可以參考如圖7所示的參考畫素範本。根據多個參考畫素生成當前待重建圖像區塊的預測訊號,可以採用上述35中預測模式中的任一種預測模式,當然也可以採用現有技術中其他的預測模式,本發明對此不做限制。In fact, which specific reconstruction image block is referenced to the reconstruction image block is related to the prediction mode adopted by the decoding end. When any of the above 35 prediction modes is used, the reference pixel template as shown in FIG. 7 can be referred to. The prediction signal of the current image block to be reconstructed is generated according to multiple reference pixels, any one of the prediction modes in 35 above may be used, and of course, other prediction modes in the prior art may also be used, the present invention does not do this limit.
若參考重建圖像區塊與當前待重建圖像區塊的解析度相同,則直接在該參考重建圖像區塊中確定至少一個參考畫素;若當前待重建圖像區塊是原解析度,參考重建圖像區塊的解析度為縮減取樣解析度,則從該參考重建圖像區塊中獲取對當前重建圖像區塊進行重建時所需的至少一個畫素,並對當前重建圖像區塊進行重建時所需的至少一個畫素進行升取樣處理,以得到當前待重建圖像區塊的至少一個參考畫素;若當前待重建圖像區塊是縮減取樣解析度,參考重建圖像區塊的解析度為原解析度,則從該參考重建圖像區塊中獲取對所述當前圖像區塊進行重建時所需的至少一個畫素,對這些畫素進行縮減取樣處理,以得到當前待重建圖像區塊的至少一個參考畫素。If the resolution of the reference image reconstruction block is the same as the current image reconstruction block, at least one reference pixel is directly determined in the reference image reconstruction block; if the current image reconstruction block is the original resolution , The resolution of the reference reconstructed image block is the downsampling resolution, then at least one pixel required for reconstruction of the current reconstructed image block is obtained from the reference reconstructed image block, and the current reconstructed image At least one pixel required for image block reconstruction is up-sampled to obtain at least one reference pixel of the current image block to be reconstructed; if the current image block to be reconstructed is down-sampled resolution, reference reconstruction If the resolution of the image block is the original resolution, at least one pixel required for reconstructing the current image block is obtained from the reference reconstructed image block, and these pixels are down-sampled To obtain at least one reference pixel of the current image block to be reconstructed.
進一步地,編碼端對當前重建圖像區塊進行的升取樣處理是基於當前重建圖像區塊所需的相鄰重建圖像區塊的畫素進行的。需要說明的是,所需的相鄰重建圖像區塊的畫素主要用於對當前重建圖像區塊的部分邊界進行升取樣處理,對於當前重建圖像區塊中除上述部分邊界之外的部分,均採用自己的畫素進行升取樣處理。假設第一濾波器是基於離散餘弦變換的插值濾波器(Discrete Cosine Transform-Based Interpolation Filter,DCTIF),這種情況下,當前重建圖像區塊所需的相鄰重建圖像區塊具體如下:圖15為本發明一實施例提供的4鄰域畫素的示意圖,如圖15所示,當前重建圖像區塊所需的相鄰重建圖像區塊包括:當前重建圖像區塊的上邊圖像區塊、下邊圖像區塊、左邊圖像區塊和右邊圖像區塊。假設第一濾波器是卷積神經網路(Convolutional Neural Network,CNN)濾波器。這種情況下,當前重建圖像區塊所需的相鄰重建圖像區塊具體如下:圖16為本發明一實施例提供的8鄰域畫素的示意圖,如圖16所示,當前重建圖像區塊所需的相鄰重建圖像區塊包括:當前重建圖像區塊的上邊圖像區塊、下邊圖像區塊、左邊圖像區塊、右邊圖像區塊、左上圖像區塊、左下圖像區塊、右上圖像區塊和右下圖像區塊。按照目前的編碼順序(從上至下,從左至右的編碼順序),目前當前重建圖像區塊的下邊圖像區塊、右邊圖像區塊、左下圖像區塊和右下圖像區塊都還沒有完成重建,現有技術是通過複製當前重建圖像區塊自己的畫素以實現升取樣處理,但是這種方式將導致經過升取樣處理後的當前重建圖像區塊存在右邊界和下邊界不連續的問題。為了解決這一問題,本發明提供如下四種可選方式:Further, the up-sampling process performed by the encoding end on the current reconstructed image block is based on the pixels of the adjacent reconstructed image block required by the current reconstructed image block. It should be noted that the required pixels of the adjacent reconstructed image block are mainly used for upsampling the partial boundary of the current reconstructed image block. For the current reconstructed image block, in addition to the above partial boundary Part of it, using its own pixels for upsampling. Assume that the first filter is a discrete cosine transform-based interpolation filter (Discrete Cosine Transform-Based Interpolation Filter, DCTIF). In this case, the neighboring reconstructed image blocks required for the current reconstructed image block are as follows: FIG. 15 is a schematic diagram of 4-neighbor pixels provided by an embodiment of the present invention. As shown in FIG. 15, the adjacent reconstructed image blocks required by the current reconstructed image block include: the upper side of the current reconstructed image block Image block, lower image block, left image block and right image block. Assume that the first filter is a Convolutional Neural Network (CNN) filter. In this case, the neighboring reconstructed image blocks required for the current reconstructed image block are specifically as follows: FIG. 16 is a schematic diagram of 8 neighbor pixels provided by an embodiment of the present invention. As shown in FIG. 16, the current reconstruction The adjacent reconstructed image blocks required for the image block include: the upper image block, the lower image block, the left image block, the right image block, the upper left image of the currently reconstructed image block Block, lower left image block, upper right image block and lower right image block. According to the current coding order (from top to bottom, from left to right coding order), the current lower image block, right image block, lower left image block and lower right image of the currently reconstructed image block The blocks have not yet been reconstructed. The existing technology is to copy the pixels of the current reconstructed image block to achieve the upsampling process, but this method will result in the right boundary of the current reconstructed image block after the upsampling process. The problem of discontinuity with the lower boundary. To solve this problem, the present invention provides the following four optional ways:
方式一、升取樣處理是在當前重建圖像區塊進行升取樣處理時所需的所有相鄰重建圖像區塊都重建完成後進行的;相應的,位元串流包括:M個參考重建圖像區塊中的每個參考重建圖像區塊的編碼方式;根據當前待重建圖像區塊的編碼方式和M個參考重建圖像區塊中的畫素,確定當前待重建圖像區塊的多個參考畫素,包括:根據當前待重建圖像區塊的編碼方式、M個參考重建圖像區塊的編碼方式和M個參考重建圖像區塊中的畫素,確定當前待重建圖像區塊的多個參考畫素。Method 1: Upsampling is performed after all the adjacent reconstructed image blocks required for the upsampling processing of the current reconstructed image block are completed; correspondingly, the bit stream includes: M reference reconstructions The encoding method of each reference reconstruction image block in the image block; according to the current encoding method of the image block to be reconstructed and the pixels in the M reference reconstruction image blocks, determine the current image area to be reconstructed Multiple reference pixels of a block, including: according to the current encoding method of the image block to be reconstructed, the encoding method of the M reference reconstruction image blocks and the pixels in the M reference reconstruction image blocks, determine the current pixel Reconstruct multiple reference pixels of an image block.
方式二、升取樣處理是在當前圖像的所有圖像區塊都重建完成後進行的;相應的,位元串流包括:M個參考重建圖像區塊中的每個參考重建圖像區塊的編碼方式;根據當前待重建圖像區塊的編碼方式和M個參考重建圖像區塊中的畫素,確定當前待重建圖像區塊的多個參考畫素,包括:根據當前待重建圖像區塊的編碼方式、M個參考重建圖像區塊的編碼方式和M個參考重建圖像區塊中的畫素,確定當前待重建圖像區塊的多個參考畫素。Method 2: Upsampling is performed after all image blocks of the current image are reconstructed; correspondingly, the bit stream includes: each reference reconstructed image area in the M reference reconstructed image blocks Block encoding method; according to the current encoding method of the image block to be reconstructed and the pixels in the M reference reconstruction image blocks, determine the multiple reference pixels of the current image block to be reconstructed, including: The encoding method of the reconstructed image block, the encoding methods of the M reference reconstructed image blocks, and the pixels in the M reference reconstructed image blocks determine multiple reference pixels of the current image block to be reconstructed.
方式三、根據所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,對當前重建圖像區塊進行一次升取樣處理;若所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Method 3: According to the pixels of a part of the neighboring reconstructed image blocks in the neighboring reconstructed image blocks that have been completely reconstructed, perform an up-sampling process on the current reconstructed image block; if required In the adjacent reconstruction image block, another part of the adjacent reconstruction image block that has not been reconstructed currently has been reconstructed, then the partial boundary of the current reconstruction image block is divided according to another part of the adjacent reconstruction image block. Sub-upsampling processing, where part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
方式四、對當前重建圖像區塊進行升取樣處理包括:根據所需的相鄰重建圖像區塊中的當前當前重建圖像區塊所處的已完成重建的一部分相鄰重建圖像區塊的畫素,對當前重建圖像區塊進行一次升取樣處理;若當前圖像的所有圖像區塊均已完成重建,則根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,另一部分相鄰重建圖像區塊是在對當前重建圖像區塊進行第一次升取樣處理時未完成重建的圖像區塊;當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Method four: Upsampling the current reconstructed image block includes: according to the required adjacent reconstructed image block, a part of the adjacent reconstructed image area where the currently reconstructed image block is located and completed The pixels of the block, up-sampling the current reconstructed image block; if all the image blocks of the current image have been reconstructed, then according to the other part of the adjacent reconstructed image block The adjacent reconstructed image block performs a second upsampling process on part of the boundary of the current reconstructed image block, where another part of the adjacent reconstructed image block is the first upsampling process on the current reconstructed image block The image block that has not been reconstructed at the time; part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
其中,所述當前重建圖像區塊的部分邊界滿足條件:在對當前重建圖像區塊進行的第一次升取樣處理中,該部分邊界所需的另一部分相鄰重建圖像區塊未完成重建。Wherein, the part of the boundary of the current reconstructed image block satisfies the condition: in the first upsampling process performed on the current reconstructed image block, another part of the adjacent reconstructed image block required by the part of the boundary is not Complete reconstruction.
可選地,當前重建圖像區塊的部分邊界為當前重建圖像區塊的右邊界和下邊界。Optionally, part of the boundary of the current reconstructed image block is the right boundary and the lower boundary of the current reconstructed image block.
可選地,所述一部分相鄰重建圖像區塊為所述當前重建圖像區塊的上邊圖像區塊和左邊圖像區塊,所述另一部分相鄰重建圖像區塊為所述當前重建圖像區塊的下邊圖像區塊和右邊圖像區塊。Optionally, the part of the adjacent reconstructed image block is the upper image block and the left image block of the current reconstructed image block, and the other part of the adjacent reconstructed image block is the The lower image block and the right image block of the currently reconstructed image block.
或者,or,
所述一部分相鄰重建圖像區塊為所述當前重建圖像區塊的左上圖像區塊,上邊圖像區塊,右上圖像區塊和左邊圖像區塊,所述另一部分相鄰重建圖像區塊為所述當前重建圖像區塊的右邊圖像區塊,左下圖像區塊,下邊圖像區塊和右下圖像區塊。The part of the adjacent reconstructed image block is an upper left image block, an upper image block, an upper right image block and a left image block of the current reconstructed image block, and the other part is adjacent The reconstructed image block is the right image block, the lower left image block, the lower image block and the lower right image block of the current reconstructed image block.
針對方式三和方式四,步驟S1402中通過第一濾波器對當前重建圖像區塊進行升取樣處理,特指對當前重建圖像區塊進行的第一次升取樣處理。For Mode 3 and Mode 4, in step S1402, the first filter is used to upsample the current reconstructed image block, specifically the first upsampling process performed on the current reconstructed image block.
針對方式一進行詳細說明:A detailed description of method one:
具體地,針對不同的濾波器,當前重建圖像區塊所需的相鄰重建圖像區塊也不同。例如:如圖15所示,當前重建圖像區塊所需的相鄰重建圖像區塊包括:當前重建圖像區塊的上邊圖像區塊、下邊圖像區塊、左邊圖像區塊和右邊圖像區塊。如圖16所示,當前重建圖像區塊所需的相鄰重建圖像區塊包括:當前重建圖像區塊的上邊圖像區塊、下邊圖像區塊、左邊圖像區塊、右邊圖像區塊、左上圖像區塊、左下圖像區塊、右上圖像區塊和右下圖像區塊。Specifically, for different filters, adjacent reconstruction image blocks required for the current reconstruction image block are also different. For example, as shown in FIG. 15, the adjacent reconstruction image blocks required for the current reconstruction image block include: the upper image block, the lower image block, and the left image block of the current reconstruction image block And the image block on the right. As shown in FIG. 16, the adjacent reconstructed image blocks required for the current reconstructed image block include: the upper image block, the lower image block, the left image block, and the right side of the current reconstructed image block Image block, upper left image block, lower left image block, upper right image block and lower right image block.
其中,可以採用現有技術的升取樣處理方法對當前重建圖像區塊進行處理。例如:圖17為本發明一實施例提供的圖像升取樣示意圖,如圖17所示,當前重建圖像區塊B所需的相鄰重建圖像區塊1、相鄰重建圖像區塊2、相鄰重建圖像區塊3和相鄰重建圖像區塊4均已完成重建。基於此,對當前重建圖像區塊B進行升取樣處理,如圖17所示,其中B中的圓圈表示升取樣前的取樣點,×表示升取樣後的取樣點。對B進行升取樣處理時,可以先對B進行水平方向上的升取樣,再對升取樣後的訊號進行垂直方向上的升取樣;或者,可以先對B進行垂直方向上的升取樣,再對升取樣後的訊號進行水平方向上的升取樣。Among them, the current up-sampling processing method can be used to process the current reconstructed image block. For example: FIG. 17 is a schematic diagram of image upsampling provided by an embodiment of the present invention. As shown in FIG. 17, the neighboring reconstructed image block 1 and the neighboring reconstructed image block required by the current reconstructed image block B 2. Both the adjacent reconstructed image block 3 and the adjacent reconstructed image block 4 have completed reconstruction. Based on this, the up-sampling process is performed on the current reconstructed image block B, as shown in FIG. 17, where the circle in B represents the sampling point before up-sampling, and × represents the sampling point after up-sampling. When up-sampling B, you can first up-sample B in the horizontal direction, and then up-sample the signal after up-sampling in the vertical direction; or, you can up-sample B in the vertical direction first, and then The up-sampled signal is up-sampled in the horizontal direction.
特別的,若當前重建圖像區塊本身是一幅圖像的邊界圖像區塊,這種情況下,即使所需的相鄰重建圖像區塊都重建完成,在進行升取樣處理時,還是需要複製當前重建圖像的畫素。例如:如圖15所示,當當前重建圖像區塊是一幅圖像的最右側的一個圖像區塊時,它的右邊圖像區塊是不存在的,因此可以對當前重建圖像區塊包括的最右側一列的畫素進行複製,以實現升取樣處理。當然,也可以採用其他方法進行升取樣處理,本發明對此不做限制。In particular, if the current reconstructed image block itself is a boundary image block of an image, in this case, even if the required adjacent reconstructed image blocks are all reconstructed, when upsampling is performed, Still need to copy the pixels of the currently reconstructed image. For example: As shown in Figure 15, when the current reconstructed image block is the rightmost image block of an image, its right image block does not exist, so the current reconstructed image The pixels in the rightmost column included in the block are copied to achieve upsampling. Of course, other methods can also be used for up-sampling, which is not limited in the present invention.
進一步地,由於當前重建圖像區塊所需的各個相鄰重建圖像區塊的編碼方式可以是縮減取樣編碼方式,也可以是原解析度編碼方式,因此在對當前重建圖像區塊進行升取樣處理時,具體分為以下兩種情況:Further, since the encoding method of each adjacent reconstructed image block required for the current reconstructed image block may be a down-sampling encoding method or an original resolution encoding method, the current reconstruction image block During the upsampling process, there are two specific cases:
1.若某相鄰重建圖像區塊的編碼方式為縮減取樣編碼方式,則可以直接根據該相鄰重建圖像區塊中的畫素對當前重建圖像區塊進行升取樣處理。1. If the encoding method of a neighboring reconstructed image block is a down-sampling encoding method, the current reconstructed image block can be up-sampled directly according to the pixels in the neighboring reconstructed image block.
2.若某相鄰重建圖像區塊的編碼方式為原解析度編碼方式,則可以獲取該相鄰重建圖像區塊的畫素中升取樣處理所需的至少一個畫素,對這些畫素進行縮減取樣處理,並根據縮減取樣處理後的至少一個畫素對當前重建圖像區塊進行升取樣處理。2. If the encoding method of a neighboring reconstructed image block is the original resolution encoding method, you can obtain at least one pixel required for upsampling in the pixels of the neighboring reconstructed image block. The pixels are down-sampled, and the current reconstructed image block is up-sampled according to at least one pixel after the down-sampling.
具體地,相鄰重建圖像區塊主要用於對當前重建圖像區塊的部分邊界進行升取樣處理(該部分邊界根據濾波器的不同而不同),例如:如圖17所示,相鄰重建圖像區塊3採用縮減取樣編碼方式,這種情況下,可以直接利用相鄰重建圖像區塊3包括的畫素對當前重建圖像區塊B的右邊界進行升取樣處理。而相鄰重建圖像區塊4採用原解析度編碼方式,則需要對相鄰重建圖像區塊4包括的升取樣處理所需的畫素進行縮減取樣處理,或者對相鄰重建圖像區塊4採用縮減取樣處理,並根據縮減取樣處理後的畫素對當前重建圖像區塊B的下邊界進行升取樣處理。其中,對相鄰重建圖像區塊4採用縮減取樣處理,具體可以是直接取圖16中被框圈出的畫素作為縮減取樣後的取樣點。或者是對相鄰重建圖像區塊4進行垂直方向上的縮減取樣處理。圖18為本發明另一實施例提供的圖像升取樣示意圖,如圖18所示,這種是8鄰域畫素的情況,解碼端對當前重建圖像區塊C取樣的升取樣處理方法與4鄰域畫素情況下的方法類似,在此不再贅述。Specifically, the adjacent reconstructed image blocks are mainly used for upsampling the part of the boundary of the current reconstructed image block (the part of the boundary differs according to the filter), for example: as shown in FIG. 17, The reconstructed image block 3 adopts the down-sampling coding method. In this case, the pixels included in the adjacent reconstructed image block 3 can be directly used to upsample the right boundary of the currently reconstructed image block B. Whereas the adjacent reconstructed image block 4 adopts the original resolution encoding method, it is necessary to perform downsampling on the pixels required for the upsampling process included in the adjacent reconstructed image block 4, or the adjacent reconstructed image area Block 4 adopts the down-sampling process, and up-samples the lower boundary of the currently reconstructed image block B according to the pixels after the down-sampling process. Among them, the down-sampling process is adopted for the adjacent reconstructed image block 4, specifically, the pixels circled in the frame in FIG. 16 may be directly taken as the sampling points after down-sampling. Or, the adjacent reconstructed image block 4 is vertically down-sampled. 18 is a schematic diagram of image upsampling provided by another embodiment of the present invention. As shown in FIG. 18, this is the case of 8-neighbor pixels, and the decoding end upsampling processing method for sampling the current reconstructed image block C It is similar to the method in the case of 4-neighbor pixels and will not be repeated here.
需要說明的是,為了避免對當前重建圖像區塊的重複升取樣處理,可以對當前重建圖像區塊進行升取樣處理後,標識當前重建圖像區塊已完成升取樣處理。或者,按照一定的規則對當前重建圖像區塊進行升取樣處理。當升取樣處理是基於4鄰域畫素的情況時,當前重建圖像區塊的下邊圖像區塊一旦完成重建,則可以對當前重建圖像區塊進行升取樣處理。當升取樣處理是基於8鄰域畫素的情況時,當前重建圖像區塊的右下圖像區塊一旦完成重建,則可以對當前重建圖像區塊進行升取樣處理。It should be noted that, in order to avoid repeated upsampling processing on the currently reconstructed image block, the upsampling processing may be performed on the current reconstructed image block to indicate that the current reconstructed image block has completed upsampling processing. Or, upsampling the current reconstructed image block according to certain rules. When the upsampling process is based on the 4-neighbor pixel, once the reconstruction of the lower image block of the current reconstructed image block is completed, the current reconstructed image block can be upsampled. When the upsampling process is based on 8-neighbor pixels, once the reconstruction of the lower right image block of the current reconstructed image block is completed, the current reconstructed image block can be upsampled.
針對方式二進行詳細說明:A detailed description of method two:
當當前圖像的所有圖像區塊都已經完成重建之後,則對於每個重建圖像區塊來講,它所需的相鄰重建圖像區塊都已經完成重建,基於此,對任一採用縮減取樣編碼的重建圖像區塊都可以進行升取樣處理。具體升取樣處理過程類似於上述方式一的方式,本發明在此不再贅述。After all the image blocks of the current image have been reconstructed, for each reconstructed image block, the adjacent reconstruction image blocks it needs have been reconstructed. Based on this, for any All the reconstructed image blocks with downsampling coding can be upsampled. The specific upsampling process is similar to the above manner 1, and the present invention will not repeat them here.
針對方式三進行詳細說明:A detailed description of method three:
在方式三中,對當前重建圖像區塊進行的升取樣處理包括兩次升取樣處理過程。第一次升取樣處理過程為:根據所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,對當前重建圖像區塊進行一次升取樣處理。第二次升取樣處理過程為:若所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,當前重建圖像區塊的所述部分邊界與另一部分相鄰重建圖像區塊鄰接。In mode three, the upsampling process performed on the current reconstructed image block includes two upsampling processes. The first upsampling process is as follows: according to the required pixels of a part of the neighboring reconstructed image blocks in the neighboring reconstructed image blocks that have been currently reconstructed, upsample the current reconstructed image block deal with. The second upsampling process is: if another part of the adjacent reconstruction image block in the required adjacent reconstruction image block that has not been currently reconstructed has been reconstructed, then the other part of the adjacent reconstruction image area is used The block performs a sub-upsampling process on a part of the boundary of the current reconstructed image block, where the part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
可選地,在對當前重建圖像區塊進行第一次升取樣處理之前,保存縮減取樣編碼方式的當前重建圖像區塊,為後續對其他待重建圖像區塊進行預測時提供參考畫素。Optionally, before performing the first upsampling process on the current reconstructed image block, the current reconstructed image block in the downsampling coding mode is saved to provide a reference picture for subsequent prediction of other image blocks to be reconstructed Prime.
具體地,假設採用從上至下、從左至右的編碼順序,對於基於4鄰域畫素或者8鄰域畫素進行升取樣處理的情況,在對當前重建圖像區塊進行第一次升取樣處理時,它的右邊圖像區塊、下邊圖像區塊、左下圖像區塊和右下圖像區塊都還沒有完成重建。這種情況下,解碼端可以複製當前重建圖像區塊包括的最右側一列或者多列的畫素,得到插值畫素。通過這些插值畫素對當前重建圖像區塊包括的右邊界進行升取樣處理。解碼端還可以複製當前重建圖像區塊包括的最下方一行或者多行的畫素,得到插值畫素。通過這些插值畫素對當前重建圖像區塊包括的下邊界進行升取樣處理。Specifically, assuming the coding order from top to bottom and left to right, for the case of upsampling processing based on 4-neighbor pixels or 8-neighbor pixels, the first During the upsampling process, its right image block, bottom image block, bottom left image block, and bottom right image block have not yet been reconstructed. In this case, the decoding end can copy the pixels on the rightmost column or columns included in the currently reconstructed image block to obtain interpolated pixels. Through these interpolation pixels, the right boundary included in the current reconstructed image block is up-sampled. The decoding end can also copy the pixels of the bottom row or rows included in the currently reconstructed image block to obtain interpolated pixels. Through these interpolated pixels, the lower boundary included in the current reconstructed image block is up-sampled.
圖19為本發明一實施例提供的圖像升取樣示意圖,如圖19所示,假設當前重建圖像區塊為圖像區塊A,在對當前重建圖像區塊A進行第二次升取樣處理時,假設目前採用DCTIF濾波器進行升取樣處理,當前重建圖像區塊A已完成第一次升取樣處理,如上面所述使用DCTIF濾波器進行升取樣處理時,需要用到左右各四個畫素,那麼在第一次升取樣處理時,當前重建圖像區塊A的右邊四列×(×表示第一次升取樣處理後的取樣點)所需的四個參考畫素均是不全的。例如,對於最右側的一列×,每個×所需要的右邊四個參考畫素都不存在。若相鄰重建圖像區塊C完成了重建,則根據相鄰重建圖像區塊C對當前重建圖像區塊的右邊界進行第二次升取樣處理。這裡的升取樣處理方法與上述升取樣處理方法相同,在此不再贅述。19 is a schematic diagram of image upsampling provided by an embodiment of the present invention. As shown in FIG. 19, assuming that the current reconstructed image block is image block A, the second reconstruction of the current reconstructed image block A is performed. In the sampling process, it is assumed that the DCTIF filter is currently used for upsampling, and the first upsampling process of the currently reconstructed image block A has been completed. As described above, when using the DCTIF filter for upsampling, you need to use the left and right Four pixels, then in the first upsampling process, the four columns on the right of the currently reconstructed image block A × (× represents the sampling point after the first upsampling process) Is incomplete. For example, for the rightmost column ×, the four reference pixels on the right required by each × do not exist. If the adjacent reconstructed image block C completes the reconstruction, the right boundary of the currently reconstructed image block is subjected to the second upsampling process according to the adjacent reconstructed image block C. The upsampling processing method here is the same as the above upsampling processing method, and will not be repeated here.
同樣地,假設當前重建圖像區塊為圖像區塊B,在對當前重建圖像區塊B進行第二次升取樣處理時,假設目前採用DCTIF濾波器進行升取樣處理,當前重建圖像區塊B已完成第一次升取樣處理,如上面所述使用DCTIF濾波器進行升取樣處理時,需要用到上下各四個畫素,那麼在第一次升取樣處理時,當前重建圖像區塊B的下方四行×(×表示第一次升取樣處理後的取樣點)所需的四個參考畫素均是不全的。例如,對於最下方的一行×,每個×所需要的下方四個參考畫素都不存在。若相鄰重建圖像區塊C完成了重建,則根據相鄰重建圖像區塊C對當前重建圖像區塊的下邊界進行第二次升取樣處理。這裡的升取樣處理方法與上述升取樣處理方法相同,在此不再贅述。Similarly, assuming that the current reconstructed image block is image block B, when performing the second upsampling process on the current reconstructed image block B, it is assumed that the DCTIF filter is currently used for upsampling processing, and the current reconstructed image Block B has completed the first upsampling process. As described above, when using the DCTIF filter for upsampling, four pixels each are required. Then, during the first upsampling process, the current reconstructed image The four rows of pixels in the lower four rows of block B (× represents the sampling point after the first upsampling process) are all incomplete. For example, for the bottom row of ×, the lower four reference pixels required by each × do not exist. If the adjacent reconstructed image block C completes the reconstruction, the second boundary of the currently reconstructed image block is subjected to the second upsampling process according to the adjacent reconstructed image block C. The upsampling processing method here is the same as the above upsampling processing method, and will not be repeated here.
圖20為本發明另一實施例提供的圖像升取樣示意圖,如圖20所示,這種是8鄰域畫素的情況,解碼端對當前重建圖像區塊A的右邊界以及當前重建圖像區塊B的下邊界採用的升取樣處理方法與4鄰域畫素情況下的方法類似,在此不再贅述。FIG. 20 is a schematic diagram of image upsampling provided by another embodiment of the present invention. As shown in FIG. 20, this is the case of 8-neighbor pixels, and the decoder ends the right boundary of the currently reconstructed image block A and the current reconstruction. The up-sampling processing method adopted at the lower boundary of the image block B is similar to the method in the case of 4-neighbor pixels, and will not be repeated here.
需要說明的是,為了避免對當前重建圖像區塊的重複升取樣處理,可以對當前重建圖像區塊完成第二次升取樣處理後,標識當前重建圖像區塊已完成升取樣處理。或者,按照一定的規則對當前重建圖像區塊進行升取樣處理。當升取樣處理是基於4鄰域畫素的情況時,當前重建圖像區塊的下邊圖像區塊一旦完成重建,則可以對當前重建圖像區塊進行第二次升取樣處理。當升取樣處理是基於8鄰域畫素的情況時,當前重建圖像區塊的右下圖像區塊一旦完成重建,則可以對當前重建圖像區塊進行第二次升取樣處理。It should be noted that, in order to avoid repeated upsampling processing on the currently reconstructed image block, after completing the second upsampling processing on the current reconstructed image block, it can be identified that the current reconstructed image block has completed upsampling processing. Or, upsampling the current reconstructed image block according to certain rules. When the upsampling process is based on the 4-neighbor pixel, once the reconstruction of the lower image block of the current reconstructed image block is completed, the second upsampling process can be performed on the current reconstructed image block. When the upsampling process is based on the 8-neighbor pixels, once the reconstruction of the lower right image block of the current reconstructed image block is completed, the second reconstructed image block can be upsampled.
針對方式四進行詳細說明:A detailed description of method four:
當當前圖像的所有圖像區塊都已經完成重建之後,則對於每個重建圖像區塊來講,它所需的相鄰重建圖像區塊都已經完成重建,基於此,對任一已完成第一次升取樣處理後的重建圖像區塊都可以進行第二次升取樣處理。具體升取樣處理過程類似於上述方式三的方式,本發明在此不再贅述。After all the image blocks of the current image have been reconstructed, for each reconstructed image block, the adjacent reconstruction image blocks it needs have been reconstructed. Based on this, for any The reconstructed image blocks after the first upsampling process has been completed can be subjected to the second upsampling process. The specific upsampling processing process is similar to the above manner three, and the present invention will not repeat them here.
可選地,在對當前重建圖像區塊進行第一次升取樣處理之前,保存縮減取樣編碼方式的當前重建圖像區塊,為後續對其他待重建圖像區塊進行預測時提供參考畫素。Optionally, before performing the first upsampling process on the current reconstructed image block, the current reconstructed image block in the downsampling coding mode is saved to provide a reference picture for subsequent prediction of other image blocks to be reconstructed Prime.
通過上述四種方式對當前重建圖像區塊進行升取樣處理時,當前重建圖像區塊的部分邊界都是通過所需的相鄰重建圖像區塊進行升取樣處理的,而現有技術中,當前重建圖像區塊的部分邊界都是通過複製當前重建圖像區塊的畫素進行升取樣處理的,因此,本發明提供的方法可以避免當前重建圖像區塊邊界不連續的問題。When upsampling the current reconstructed image block in the above four ways, part of the boundary of the current reconstructed image block is upsampled through the required neighboring reconstructed image block, and in the prior art Part of the boundary of the current reconstructed image block is up-sampled by copying the pixels of the current reconstructed image block. Therefore, the method provided by the present invention can avoid the problem of discontinuous boundary of the current reconstructed image block.
綜所上述,相對於現有技術中對整幅圖像中的重建圖像區塊都採用相同的濾波器,本發明針對每個重建圖像區塊都要選擇對應的濾波器,即本發明採用具有針對性的選擇方式來選擇濾波器,通過選擇的濾波器對重建圖像區塊進行升取樣處理,從而可以得到顯示效果更佳的重建圖像區塊。In summary, relative to the prior art, which uses the same filter for the reconstructed image blocks in the entire image, the present invention selects the corresponding filter for each reconstructed image block, that is, the present invention uses The filter is selected in a targeted selection manner, and the reconstructed image block is up-sampled by the selected filter, so that a reconstructed image block with better display effect can be obtained.
可選地,可以通過如下幾種情況選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器:Optionally, the first filter used for upsampling the current reconstructed image block can be selected in the following situations:
具體地,情況一:從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器具體包括:當前重建圖像區塊的至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊與當前重建圖像區塊的相似度,其中,至少兩個相鄰重建圖像區塊進行升取樣處理時對應至少兩個第二濾波器,至少兩個候選濾波器包括至少兩個第二濾波器;從至少兩個第二濾波器中選擇與當前重建圖像區塊相似度最高的相鄰重建圖像區塊對應的第二濾波器作為第一濾波器。Specifically, case 1: selecting the first filter for upsampling the current reconstructed image block from at least two candidate filters specifically includes: at least two adjacent reconstruction images of the current reconstructed image block The similarity between each adjacent reconstructed image block in the image block and the current reconstructed image block, where at least two adjacent reconstructed image blocks correspond to at least two second filters when upsampling is performed, The at least two candidate filters include at least two second filters; from the at least two second filters, the second filter corresponding to the adjacent reconstructed image block with the highest similarity of the current reconstructed image block is selected as the second filter First filter.
其中,將第二濾波器作為第一濾波器不僅僅表示為一種賦值關係,還表示直接選擇第二濾波器對當前重建圖像區塊進行升取樣處理。Among them, the second filter as the first filter is not only represented as an assignment relationship, but also means that the second filter is directly selected to upsample the current reconstructed image block.
針對上述方式一和方式二,相鄰重建圖像區塊是指:在當前重建圖像區塊的所有相鄰重建圖像區塊中,已經完成升取樣處理的重建圖像區塊。針對上述方式三和方式四,相鄰重建圖像區塊是指:在當前重建圖像區塊的所有相鄰重建圖像區塊中,已經完成第一次升取樣處理或者第二次升取樣處理的重建圖像區塊。針對方式三和方式四,第二濾波器用於對相鄰重建圖像區塊進行的第一次升取樣處理,或者,第二濾波器用於對相鄰重建圖像區塊進行的第二次升取樣處理,本發明對此不做限制。With respect to the first and second manners above, the neighboring reconstructed image blocks refer to: among all the neighboring reconstructed image blocks of the current reconstructed image block, the reconstructed image blocks that have completed the upsampling process. Regarding the above method 3 and method 4, the adjacent reconstructed image block refers to: in all adjacent reconstructed image blocks of the current reconstructed image block, the first upsampling process or the second upsampling process has been completed Processed reconstructed image block. For Mode 3 and Mode 4, the second filter is used for the first upsampling process of adjacent reconstructed image blocks, or the second filter is used for the second upsampling process of adjacent reconstructed image blocks The sampling process is not limited by the present invention.
計算為當前重建圖像區塊與相鄰重建圖像區塊的相似度的方法可以是:若當前重建圖像區塊與相鄰重建圖像區塊的解析度相同,則對當前重建圖像區塊的每個畫素與相鄰重建圖像區塊中對應的畫素求差,得到對應差值,對當前重建圖像區塊的所有差值求加權平均值,最終得到當前重建圖像區塊與相鄰重建圖像區塊的誤差,該誤差越小,則表示相似度越高。若當前重建圖像區塊與相鄰重建圖像區塊的解析度不同,則先對相鄰重建圖像區塊進行取樣處理,使得取樣處理後的相鄰重建圖像區塊的解析度與當前重建圖像區塊的解析度相同,然後對當前重建圖像區塊的每個畫素與取樣處理後的相鄰重建圖像區塊中對應的畫素求差,得到對應差值,對當前重建圖像區塊的所有差值求加權平均值,最終得到當前重建圖像區塊與相鄰重建圖像區塊的誤差,該誤差越小,則表示相似度越高。The method for calculating the similarity between the current reconstructed image block and the adjacent reconstructed image block may be: if the resolution of the current reconstructed image block and the adjacent reconstructed image block are the same, the current reconstructed image Each pixel of the block is compared with the corresponding pixel in the adjacent reconstructed image block to obtain the corresponding difference value, and the weighted average of all the difference values of the current reconstructed image block is finally obtained to finally obtain the current reconstructed image The error between a block and an adjacent reconstructed image block. The smaller the error, the higher the similarity. If the resolution of the current reconstructed image block and the neighboring reconstructed image block are different, the neighboring reconstructed image block is sampled first, so that the resolution of the neighboring reconstructed image block after the sampling process is The resolution of the current reconstructed image block is the same, and then each pixel of the current reconstructed image block is subtracted from the corresponding pixel in the adjacent reconstructed image block after sampling to obtain the corresponding difference value. The weighted average of all differences in the current reconstructed image block finally obtains the error between the current reconstructed image block and the adjacent reconstructed image block. The smaller the error, the higher the similarity.
需要說明的是,本發明對確定當前重建圖像區塊與相鄰重建圖像區塊的相似度不限於此。It should be noted that the present invention is not limited to determining the similarity between the current reconstructed image block and the adjacent reconstructed image block.
進一步地,若當前重建圖像區塊不存在上述相鄰重建圖像區塊,即針對上述方式一和方式二,在當前重建圖像區塊的所有相鄰重建圖像區塊中,不存在已經完成升取樣處理的重建圖像區塊。或者,針對上述方式三和方式四,在當前重建圖像區塊的所有相鄰重建圖像區塊中,不存在已經完成第一次升取樣處理和第二次升取樣處理的重建圖像區塊。則可以計算這些相鄰重建圖像區塊的相鄰重建圖像區塊與當前重建圖像區塊的相似度。相鄰重建圖像區塊的相鄰重建圖像區塊進行升取樣處理時對應至少兩個濾波器;從至少兩個濾波器中選擇與當前重建圖像區塊相似度最高的相鄰重建圖像區塊的相鄰重建圖像區塊對應的濾波器作為第一濾波器。本發明對此不做限制。Further, if the above-mentioned adjacent reconstructed image blocks do not exist in the current reconstructed image block, that is to say, in the above manners 1 and 2, there is no existing among all adjacent reconstructed image blocks in the current reconstructed image block The reconstructed image block that has completed the upsampling process. Or, with respect to the above method 3 and method 4, in all adjacent reconstructed image blocks of the current reconstructed image block, there is no reconstructed image area that has completed the first upsampling process and the second upsampling process Piece. Then, the similarity between the adjacent reconstructed image block and the current reconstructed image block of these adjacent reconstructed image blocks can be calculated. The adjacent reconstructed image blocks of the adjacent reconstructed image blocks correspond to at least two filters when performing upsampling processing; the adjacent reconstructed image with the highest similarity to the current reconstructed image block is selected from the at least two filters The filter corresponding to the adjacent reconstructed image block of the image block serves as the first filter. The present invention does not limit this.
可選地,本發明還可以將當前重建圖像區塊的任一個相鄰重建圖像區塊進行升取樣處理時對應的第二濾波器作為第一濾波器。若當前重建圖像區塊不存在上述相鄰重建圖像區塊,即針對上述方式一和方式二,在當前重建圖像區塊的所有相鄰重建圖像區塊中,不存在已經完成升取樣處理的重建圖像區塊。或者,針對上述方式三和方式四,在當前重建圖像區塊的所有相鄰重建圖像區塊中,不存在已經完成第一次升取樣處理和第二次升取樣處理的重建圖像區塊。則可以選擇這些相鄰重建圖像區塊的所有相鄰重建圖像區塊中需要進行升取樣處理的相鄰重建圖像區塊。最後,將選擇的相鄰重建圖像區塊進行升取樣處理時對應的濾波器作為第一濾波器本發明對此不做限制。Optionally, the present invention may also use the second filter corresponding to any adjacent reconstructed image block of the current reconstructed image block as the first filter when performing upsampling processing. If the current reconstructed image block does not have the above-mentioned adjacent reconstructed image blocks, that is, for the first and second ways above, among all the adjacent reconstructed image blocks of the current reconstructed image block, there is no already completed upgrade. Reconstructed image block from sampling process. Or, with respect to the above method 3 and method 4, in all adjacent reconstructed image blocks of the current reconstructed image block, there is no reconstructed image area that has completed the first upsampling process and the second upsampling process Piece. Then, it is possible to select the adjacent reconstructed image blocks that need to be up-sampled among all the adjacent reconstructed image blocks of these adjacent reconstructed image blocks. Finally, the corresponding filter when the selected adjacent reconstructed image block is up-sampled is used as the first filter. The present invention does not limit this.
情況二、從至少兩個候選濾波器中選擇用於對所述當前重建圖像區塊進行升取樣處理的第一濾波器包括:確定對當前重建圖像區塊的至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊進行升取樣處理時使用的至少兩個第二濾波器;至少兩個候選濾波器包括所述至少兩個第二濾波器,按照各個相鄰重建圖像區塊的編號順序從至少兩個第二濾波器中選擇第一個相鄰重建圖像區塊對應的第二濾波器,作為第一濾波器。Case 2: The first filter selected from at least two candidate filters for upsampling the current reconstructed image block includes: determining at least two adjacent reconstructed images of the current reconstructed image block At least two second filters used in the upsampling process of each adjacent reconstructed image block in the image block; at least two candidate filters include the at least two second filters, reconstructed according to each adjacent The numbering order of the image blocks selects the second filter corresponding to the first adjacent reconstructed image block from the at least two second filters as the first filter.
例如:圖21為本發明一實施例提供的當前重建圖像區塊和相鄰重建圖像區塊的示意圖,如圖21所示,判斷相鄰重建圖像區塊A0、A1、B0、B1、B2中已經完成升取樣處理的相鄰重建圖像區塊,假設A0、A1、B0、B1、B2都是相鄰重建圖像區塊,並且它們的編碼順序是:B1,A1,B2,B0,A0,按照編碼順序,則選擇B1對應的第二濾波器作為當前重建圖像區塊的第一濾波器。For example: FIG. 21 is a schematic diagram of a current reconstructed image block and an adjacent reconstructed image block provided by an embodiment of the present invention. As shown in FIG. 21, the adjacent reconstructed image blocks A0, A1, B0, and B1 are determined. , The adjacent reconstructed image blocks in B2 that have been upsampled, assuming that A0, A1, B0, B1, and B2 are all adjacent reconstructed image blocks, and their coding order is: B1, A1, B2, B0, A0, according to the coding order, the second filter corresponding to B1 is selected as the first filter of the current reconstructed image block.
情況三、從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器包括:確定對當前重建圖像區塊的至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊進行升取樣處理時使用的至少兩個第二濾波器;至少兩個候選濾波器包括所述至少兩個第二濾波器,選擇使用概率最高的第二濾波器作為第一濾波器。Case 3: The first filter selected from at least two candidate filters for upsampling the current reconstructed image block includes: determining at least two adjacent reconstructed image areas of the current reconstructed image block At least two second filters used in the upsampling process of each adjacent reconstructed image block in the block; at least two candidate filters include the at least two second filters, and the second with the highest probability of use is selected The filter serves as the first filter.
情況四、從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器包括:根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器中選擇第一濾波器。其中,根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器中選擇第一濾波器,包括:根據預設映射關係和當前重建圖像區塊的紋理特徵選擇第一濾波器,預設映射關係是包括當前重建圖像區塊的紋理特徵的預設紋理特徵和包括第一濾波器的所述至少兩個候選濾波器之間的映射關係。Case four: Selecting the first filter for upsampling the current reconstructed image block from at least two candidate filters includes: selecting from at least two candidate filters according to the texture characteristics of the current reconstructed image block Select the first filter. Wherein, selecting the first filter from the at least two candidate filters according to the texture feature of the current reconstructed image block includes: selecting the first filter according to the preset mapping relationship and the texture feature of the current reconstructed image block, Let the mapping relationship be the mapping relationship between the preset texture features including the texture features of the current reconstructed image block and the at least two candidate filters including the first filter.
具體地,檢測當前重建圖像區塊的紋理特徵包括:邊緣檢測方式或者在頻域中判斷紋理特徵的方式等。下面以第二種方式為例:Specifically, detecting the texture feature of the currently reconstructed image block includes: an edge detection method or a method for determining the texture feature in the frequency domain. The following takes the second method as an example:
圖22為本發明一實施例提供的當前重建圖像區塊的示意圖,如圖22所示,該當前重建圖像區塊是一個解析度為8*8的圖像區塊。當對當前重建圖像區塊經過離散餘弦變換(Discrete Cosine Transform,DCT)變換之後,會生成如圖22所示的0~63個係數,橫向用i表示,i=0~7;縱向用j表示,j=0~7。FIG. 22 is a schematic diagram of a current reconstructed image block provided by an embodiment of the present invention. As shown in FIG. 22, the current reconstructed image block is an image block with a resolution of 8 * 8. After the discrete cosine transform (Discrete Cosine Transform, DCT) transformation is performed on the current reconstructed image block, 0 to 63 coefficients as shown in FIG. 22 are generated, i is represented in the horizontal direction, i = 0 to 7; j is used in the vertical direction Indicates that j = 0 ~ 7.
當當前重建圖像區塊在頻域上滿足公式,則表示該當前重建圖像區塊的紋理特徵為平坦,否則,則表示該當前重建圖像區塊的紋理特徵為紋理。When the current reconstructed image block satisfies the formula in the frequency domain , It means that the texture feature of the current reconstructed image block is flat, otherwise, it means that the texture feature of the current reconstructed image block is texture.
其中,上述公式左側表示當前重建圖像區塊中所有畫素的交流電流(Alternating Current,AC)的平方和,右側為當前重建圖像區塊中所有畫素的直流電流(Direct Current t,DC)的平方和與係數a之積,其中,a大於0且小於等於1。例如:a可以取經驗值0.02。Among them, the left side of the above formula represents the sum of squares of alternating currents (Alternating Current, AC) of all pixels in the currently reconstructed image block, and the right side is the direct currents (Direct Current t, DC) of all pixels in the currently reconstructed image block. ) Is the product of the sum of squares and the coefficient a, where a is greater than 0 and less than or equal to 1. For example: a can take the experience value of 0.02.
進一步地,紋理特徵和濾波器之間的預設映射關係可以是表1所示的映射關係,本發明對預設映射關係不做限制。Further, the preset mapping relationship between the texture feature and the filter may be the mapping relationship shown in Table 1, and the present invention does not limit the preset mapping relationship.
表1
最後,可以根據上述預設映射關係和當前重建圖像區塊的紋理特徵選擇第一濾波器。假設當前重建圖像區塊的紋理特徵是平臺,預設映射關係如表1,則選擇的第一濾波器為DCTIF濾波器。Finally, the first filter can be selected according to the above-mentioned preset mapping relationship and the texture feature of the currently reconstructed image block. Assuming that the texture feature of the currently reconstructed image block is a platform and the preset mapping relationship is as shown in Table 1, the first filter selected is a DCTIF filter.
情況五、從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器包括:根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器中選擇第一濾波器。其中,根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器中選擇第一濾波器,包括:在所有相鄰重建圖像區塊中確定與當前重建圖像區塊紋理特徵相同的至少兩個相鄰重建圖像區塊,至少兩個相鄰重建圖像區塊進行升取樣處理時對應至少兩個第二濾波器,至少兩個候選濾波器包括至少兩個第二濾波器,從所述至少兩個第二濾波器中選擇與當前重建圖像區塊紋理特徵相同的相鄰重建圖像區塊對應的第二濾波器作為第一濾波器。Case 5: Selecting the first filter for upsampling the current reconstructed image block from at least two candidate filters includes: selecting from at least two candidate filters according to the texture characteristics of the current reconstructed image block Select the first filter. Wherein, selecting the first filter from at least two candidate filters according to the texture features of the current reconstructed image block includes: determining, in all adjacent reconstructed image blocks, the same texture features as the current reconstructed image block At least two adjacent reconstructed image blocks, the at least two adjacent reconstructed image blocks corresponding to at least two second filters when upsampling is processed, and the at least two candidate filters include at least two second filters, A second filter corresponding to an adjacent reconstructed image block having the same texture feature as the current reconstructed image block from the at least two second filters is selected as the first filter.
其中,相鄰重建圖像區塊的紋理特徵與第二濾波器的映射關係可以參考上述表1中的預設映射關係。本發明對此不做限制。The mapping relationship between the texture feature of the adjacent reconstructed image block and the second filter can refer to the preset mapping relationship in Table 1 above. The present invention does not limit this.
情況六、從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,包括:通過至少兩個候選濾波器分別對當前重建圖像區塊進行升取樣處理,得到至少兩個候選濾波器分別對應的升取樣圖像區塊;分別計算至少兩個候選濾波器對應的升取樣圖像區塊與當前重建圖像區塊對應的原始圖像區塊的誤差;將最小誤差對應的候選濾波器作為第一濾波器。Case 6: Selecting a first filter for upsampling the current reconstructed image block from at least two candidate filters, including: upsampling the current reconstructed image block through at least two candidate filters respectively Sampling process to obtain upsampled image blocks corresponding to at least two candidate filters respectively; calculate the upsampled image blocks corresponding to at least two candidate filters and the original image blocks corresponding to the current reconstructed image block Error; the candidate filter corresponding to the smallest error is used as the first filter.
其中,通過濾波器對當前重建圖像區塊進行升取樣處理的方法可以參考上述升取樣過程,在此不再贅述。計算升取樣圖像區塊與原始圖像區塊的誤差,包括:對升取樣圖像區塊的每個畫素與原始圖像區塊中對應的畫素求差,得到對應差值,對所有差值求加權平均值,最終得到升取樣圖像區塊與原始圖像區塊的誤差。其中,升取樣圖像區塊與原始圖像區塊的誤差越小,表示升取樣圖像區塊對應濾波器的升取樣處理效果越好。相反,升取樣圖像區塊與原始圖像區塊的誤差越大,表示升取樣圖像區塊對應濾波器的升取樣處理效果越差。The method of upsampling the current reconstructed image block through the filter can refer to the above upsampling process, which will not be repeated here. Calculate the error between the upsampled image block and the original image block, including: taking the difference between each pixel of the upsampled image block and the corresponding pixel in the original image block to obtain the corresponding difference, right The weighted average of all differences is obtained, and finally the error between the upsampled image block and the original image block is obtained. The smaller the error between the up-sampled image block and the original image block, the better the up-sampling processing effect of the corresponding filter of the up-sampled image block. Conversely, the greater the error between the upsampled image block and the original image block, the worse the upsampling processing effect of the corresponding filter of the upsampled image block.
需要說明的是,本發明中還可以根據亮度分量和色度分量選擇濾波器。It should be noted that, in the present invention, the filter may also be selected according to the luminance component and the chrominance component.
綜所上述,本發明可以根據上述六種情況為當前重建圖像區塊選擇第一濾波器。相對於現有技術中對整幅圖像中的重建圖像區塊都採用相同的濾波器,本發明針對每個重建圖像區塊都要選擇對應的濾波器,從而可以得到顯示效果更佳的重建圖像區塊。In summary, the present invention can select the first filter for the current reconstructed image block according to the above six situations. Compared with the prior art that uses the same filter for the reconstructed image blocks in the entire image, the present invention must select a corresponding filter for each reconstructed image block, so that a better display effect can be obtained Rebuild the image block.
當對當前重建圖像區塊採用上述方式三或者方式四進行升取樣處理時,由於存在第二次升取樣處理,因此,需要通過第三濾波器對當前重建圖像區塊進行第二次升取樣處理。When the current reconstructed image block is upsampled using the above method 3 or method 4, since there is a second upsampling process, a third filter is needed to upsample the current reconstructed image block Sampling processing.
本發明提供一種圖像處理方法,其中第三濾波器可以是上述的第一濾波器,或者可以參照上述選擇第一濾波器的方式選擇第三濾波器,需要強調的是,該第三濾波器是針對當前重建圖像區塊的部分邊界進行升取樣處理的。The present invention provides an image processing method, wherein the third filter may be the above-mentioned first filter, or the third filter may be selected with reference to the above-mentioned way of selecting the first filter, and it should be emphasized that the third filter Upsampling is performed on part of the boundary of the currently reconstructed image block.
可選地,通過第三濾波器,根據在所述所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理之前,還包括:根據另一部分相鄰重建圖像區塊和當前重建圖像區塊的部分邊界判斷是否對部分邊界進行二次升取樣處理,若確定對第一重建圖像區塊的部分邊界進行二次升取樣處理,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, through a third filter, the partial boundary of the current reconstructed image block is quadratic according to another partial adjacent reconstructed image block in the desired adjacent reconstructed image block Before the upsampling process, it also includes: judging whether to perform a second upsampling process on the part of the boundary according to another part of the adjacent reconstructed image block and the part of the boundary of the current reconstructed image block, if it is determined that the first reconstructed image block The second part of the boundary is subjected to sub-upsampling, and then the third filter is used to perform sub-upsampling on the part of the boundary of the current reconstructed image block according to another part of the adjacent reconstructed image block.
具體地,若當前重建圖像區塊的部分邊界包括:它的右邊界和下邊界。則在另一部分第二相鄰重建圖像區塊中確定與右邊界鄰接的一列或者多列畫素。根據經過第一次升取樣處理後的右邊界與該一列或者多列的畫素(構成右邊界的鄰接邊界)確定是否對右邊界進行二次升取樣處理,當確定對右邊界進行二次升取樣處理時,也認為對下邊界進行二次升取樣處理。Specifically, if a part of the boundary of the current reconstructed image block includes: its right boundary and lower boundary. Then, one or more columns of pixels adjacent to the right boundary are determined in another part of the second adjacent reconstructed image block. According to the right border after the first upsampling process and the pixels in one or more columns (the adjacent borders that constitute the right border), determine whether to perform the second upsampling process on the right border. In the sampling process, it is also considered that the lower boundary is double-upsampled.
或者,or,
若當前重建圖像區塊的部分邊界包括:它的右邊界和下邊界。則在另一部分第二相鄰重建圖像區塊中確定與下邊界鄰接的一行或者多行畫素。根據經過第一次升取樣處理後的下邊界與該一行或者多行的畫素(構成下邊界的鄰接邊界)確定是否對下邊界進行二次升取樣處理,當確定對下邊界進行二次升取樣處理時,也認為對右邊界進行二次升取樣處理。If the part of the boundary of the current reconstructed image block includes: its right boundary and lower boundary. Then, one or more lines of pixels adjacent to the lower boundary are determined in another part of the second adjacent reconstructed image block. According to the lower boundary after the first upsampling process and the pixels of one or more lines (adjacent borders constituting the lower boundary), determine whether to perform the second upsampling process on the lower boundary. When sampling, it is also considered that the right boundary is upsampled.
其中,根據經過第一次升取樣處理後的右邊界與該一列或者多列的畫素確定是否對右邊界進行二次升取樣處理,具體如下:According to the right boundary after the first upsampling process and the pixels in the column or columns, it is determined whether to perform the second upsampling process on the right boundary, as follows:
圖23為本發明一實施例提供的右邊界和右邊界的鄰接邊界的示意圖,如圖23所示,23 is a schematic diagram of a right border and an adjacent border of a right border provided by an embodiment of the present invention, as shown in FIG. 23,
如果滿足下面所有條件時,則確定對右邊界進行二次升取樣處理,否則,不對右邊界進行二次升取樣處理。If all the following conditions are met, it is determined that the right boundary is upsampled, otherwise, the right boundary is not upsampled.
|p0–q0|<TH1| p0–q0 | < TH1
|p1–p0|<TH2| p1–p0 | < TH2
|q1–q0|<TH3| q1–q0 | < TH3
其中,p0至p3表示右邊界的各個畫素的畫素值,q0至q3表示右邊界的鄰接邊界的各個畫素的畫素值。TH1、TH2、TH3分別是預設閾值,它們可以相同,也可以不同。Among them, p0 to p3 represent the pixel value of each pixel of the right boundary, and q0 to q3 represent the pixel value of each pixel of the adjacent boundary of the right boundary. TH1, TH2, and TH3 are preset thresholds, respectively, and they may be the same or different.
通過上述方法可以有效確定是否對部分邊界進行二次升取樣處理。Through the above method, it is possible to effectively determine whether to partially re-upsample some boundaries.
可選地,上述方法還包括:生成位元串流,該位元串流包括:第一濾波器的標識資訊。Optionally, the above method further includes: generating a bit stream, the bit stream includes: identification information of the first filter.
其中,可以實現升取樣處理的濾波器一般為一個或多個,它們可以是不同類型,不同抽頭數的濾波器。也可以是相同類型,不同抽頭數的濾波器。又或者是相同類型,相同抽頭數,不同係數的濾波器。其中,每個濾波器都具有對應的標識資訊。具體見表2、表3和表4。Among them, there are generally one or more filters that can implement upsampling, and they can be filters of different types and different numbers of taps. It can also be a filter of the same type and different numbers of taps. Or filters of the same type, the same number of taps, and different coefficients. Among them, each filter has corresponding identification information. See Table 2, Table 3 and Table 4 for details.
表2:
表3:
表4:
通過攜帶第一濾波器的標識資訊的方式,使得解碼端可以通過第一濾波器對當前重建圖像區塊進行升取樣處理。By carrying the identification information of the first filter, the decoding end can upsample the current reconstructed image block through the first filter.
可選地,生成位元串流,該位元串流包括:第一指示資訊,其中,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器。Optionally, a bit stream is generated, the bit stream includes: first indication information, wherein the first indication information is used to indicate how to upsample the currently reconstructed image block from at least two candidate filters Filter used in processing.
可選地,對濾波器的選擇方式包括:方式一、根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;方式二、根據當前重建圖像區塊的相鄰重建圖像區塊從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器。Optionally, the selection method of the filter includes: method 1: selecting a filter used for upsampling the current reconstructed image block from at least two candidate filters according to the texture characteristics of the current reconstructed image block; Manner 2: The filter used when upsampling the current reconstructed image block is selected from at least two candidate filters according to adjacent reconstructed image blocks of the current reconstructed image block.
可選地,生成位元串流,該位元串流包括:第二指示資訊,其中,第二指示資訊用於指示解碼端是否需要對當前重建圖像區塊進行二次升取樣處理。Optionally, a bit stream is generated, the bit stream includes: second indication information, wherein the second indication information is used to indicate whether the decoding end needs to perform up-sampling processing on the current reconstructed image block.
需要說明的是,上述位元串流可以包括以下至少一項:第一濾波器的標識資訊、第一指示資訊、第二指示資訊等,可選地,該位元串流還包括:當前待編碼圖像區塊的編碼資訊等,解碼端可以通過對該編碼資訊進行亂度解碼得到當前待編碼圖像區塊的變換量化係數,還可以通過該編碼資訊得到預測訊號等。該位元串流還可以包括:當前待編碼圖像區塊的編碼方式等。It should be noted that the above bit stream may include at least one of the following: identification information of the first filter, first indication information, second indication information, etc. Optionally, the bit stream further includes: the current pending For the coding information of the coded image block, the decoding end can obtain the transform quantization coefficient of the current image block to be coded by scrambling the coded information, and can also obtain the prediction signal by using the coding information. The bit stream may also include: the current encoding method of the image block to be encoded.
上面主要介紹了編碼端的圖像處理方法,下面將介紹解碼端的圖像處理方法。The above mainly introduces the image processing method of the encoding end, and the image processing method of the decoding end will be introduced below.
具體地,圖24為本發明另一實施例提供的一種圖像處理方法的流程圖,如圖24所示,該方法包括:Specifically, FIG. 24 is a flowchart of an image processing method according to another embodiment of the present invention. As shown in FIG. 24, the method includes:
步驟S2401:解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和當前待重建圖像區塊的編碼方式;Step S2401: Parse the bit stream to obtain the encoding information of the image block to be reconstructed and the encoding method of the image block to be reconstructed;
步驟S2402:根據當前待重建圖像區塊的編碼資訊生成當前待重建圖像區塊的重建訊號;Step S2402: generate a reconstruction signal of the current image block to be reconstructed according to the coding information of the current image block to be reconstructed;
步驟S2403:根據重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;Step S2403: Reconstruct the current image block to be reconstructed according to the reconstruction signal to obtain the current reconstructed image block;
步驟S2404:若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則根據從位元串流獲取的第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器;Step S2404: If the current encoding method of the image block to be reconstructed is a down-sampling encoding method, then select the first for upsampling the current reconstructed image block according to the first indication information obtained from the bit stream filter;
步驟S2405:通過第一濾波器對當前重建圖像區塊進行升取樣處理。Step S2405: Upsampling the current reconstructed image block through the first filter.
其中,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器。The first indication information is used to indicate how to select the filter used for upsampling the current reconstructed image block from at least two candidate filters.
該位元串流還包括:當前待重建圖像區塊的編碼方式、當前待重建圖像區塊項所述之當前圖像中已完成重建的重建圖像區塊的編碼方式。或者,該位元串流包括一些編碼參數,通過這些編碼參數可以確定當前待重建圖像區塊的編碼方式、當前待重建圖像區塊項所述之當前圖像中已完成重建的重建圖像區塊的編碼方式等。The bit stream also includes: the encoding method of the current image block to be reconstructed, and the encoding method of the reconstructed image block in the current image described in the item of the current image block to be reconstructed. Or, the bit stream includes some coding parameters, which can be used to determine the encoding method of the current image block to be reconstructed, the reconstruction image of the current image that has been reconstructed in the current image block item to be reconstructed Like the encoding method of the block.
該編碼資訊用於生成待重建圖像區塊的重建訊號,比如:解碼端通過對編碼資訊進行亂度解碼,得到當前待重建圖像區塊的變換量化係數,然後對變換量化係數進行反量化和反變換,得到當前待重建圖像區塊的重建的殘差訊號。通過它的參考重建圖像區塊(參考重建圖像區塊的資訊屬於編碼資訊)對當前待重建圖像區塊進行預測,得到當前待重建圖像區塊的預測訊號,然後將預測訊號和上述重建的殘差訊號相加,得到當前待重建圖像區塊的重建訊號。該位元串流還包括現有技術涉及的更多的編碼資訊,本發明對此不做限制。The coding information is used to generate reconstruction signals of the image block to be reconstructed. For example, the decoding end obtains the transform and quantization coefficients of the current image block to be reconstructed by random decoding the encoded information, and then inversely quantizes the transform and quantization coefficients And inverse transform to obtain the reconstructed residual signal of the current image block to be reconstructed. Through its reference reconstruction image block (the information of the reference reconstruction image block belongs to the coding information), the current image block to be reconstructed is predicted to obtain the prediction signal of the current image block to be reconstructed, and then the prediction signal and The reconstructed residual signals are added to obtain the reconstruction signal of the current image block to be reconstructed. The bit stream also includes more coding information related to the prior art, which is not limited by the present invention.
可選地,對濾波器的選擇方式包括:方式一、根據所述當前重建圖像區塊的紋理特徵從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;方式二、根據當前重建圖像區塊的相鄰重建圖像區塊從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器。Optionally, the selection method of the filter includes: method one. Select the filter used for upsampling the current reconstructed image block from at least two candidate filters according to the texture characteristics of the current reconstructed image block Method 2: Select the filter used for upsampling the current reconstructed image block from at least two candidate filters according to adjacent reconstructed image blocks of the current reconstructed image block.
最後,若當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則根據第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,通過第一濾波器對當前重建圖像區塊進行升取樣處理。Finally, if the encoding method of the current reconstructed image block is a down-sampling encoding method, the first filter used for upsampling the current reconstructed image block is selected according to the first instruction information, and the first filter is used to The current reconstructed image block is up-sampled.
需要說明的是,解碼端和編碼端涉及的重建當前待重建圖像區塊方法類似,並且解碼端和編碼端涉及的對當前重建圖像區塊進行升取樣處理地方法也類似。在此不再贅述這兩個方法。It should be noted that the decoding end and the encoding end are involved in a method of reconstructing the current image block to be reconstructed, and the decoding end and the encoding end are involved in a method of upsampling the currently reconstructed image block. These two methods are not repeated here.
綜所上述,相對於現有技術中對整幅圖像中的重建圖像區塊都採用相同的濾波器,本發明針對每個重建圖像區塊都要選擇對應的濾波器,即本發明採用具有針對性的選擇方式來選擇濾波器,通過選擇的濾波器對重建圖像區塊進行升取樣處理,從而可以得到顯示效果更佳的重建圖像區塊。In summary, relative to the prior art, which uses the same filter for the reconstructed image blocks in the entire image, the present invention selects the corresponding filter for each reconstructed image block, that is, the present invention uses The filter is selected in a targeted selection manner, and the reconstructed image block is up-sampled by the selected filter, so that a reconstructed image block with better display effect can be obtained.
可選地,第一指示資訊用於指示根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;根據從位元串流獲取的第一指示資訊選擇用於對所述當前重建圖像區塊進行升取樣處理的第一濾波器,包括:根據預設映射關係和當前重建圖像區塊的紋理特徵選擇第一濾波器,預設映射關係是包括當前重建圖像區塊的紋理特徵的預設紋理特徵和包括第一濾波器的至少兩個候選濾波器之間的映射關係。Optionally, the first indication information is used to instruct the selection of the filter used for upsampling the current reconstructed image block from at least two candidate filters according to the texture characteristics of the current reconstructed image block; according to the slave bit The first indication information acquired by the stream selects the first filter for upsampling the current reconstructed image block, including: selecting the first filter according to the preset mapping relationship and the texture feature of the current reconstructed image block For the filter, the preset mapping relationship is a mapping relationship between the preset texture feature including the texture feature of the currently reconstructed image block and at least two candidate filters including the first filter.
可選地,第一指示資訊用於指示根據當前重建圖像區塊的相鄰重建圖像區塊從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;根據從位元串流獲取的第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,包括:確定至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊與所述當前重建圖像區塊的相似度,至少兩個相鄰重建圖像區塊進行升取樣處理時對應至少兩個第二濾波器,至少兩個候選濾波器包括至少兩個第二濾波器;從至少兩個第二濾波器中選擇相似度最高的相鄰重建圖像區塊對應的第二濾波器作為所述第一濾波器。Optionally, the first indication information is used to instruct the filter used for upsampling the current reconstructed image block to be selected from at least two candidate filters according to adjacent reconstructed image blocks of the current reconstructed image block Selecting the first filter for upsampling the current reconstructed image block according to the first indication information obtained from the bit stream, including: determining each phase in at least two adjacent reconstructed image blocks The similarity between the adjacent reconstructed image block and the current reconstructed image block, at least two second filters corresponding to at least two adjacent reconstructed image blocks when upsampling is processed, and the at least two candidate filters include At least two second filters; selecting the second filter corresponding to the adjacent reconstructed image block with the highest similarity from the at least two second filters as the first filter.
可選地,通過第一濾波器對當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對當前重建圖像區塊進行一次升取樣處理;相應的,所述方法還包括:若所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: the currently completed neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block A part of the reconstructed pixels of adjacent reconstructed image blocks are subjected to an up-sampling process on the current reconstructed image block through the first filter; correspondingly, the method further includes: If another part of the adjacent reconstruction image block in the block that has not been reconstructed currently has been reconstructed, the third filter is used to perform partial boundary reconstruction of the current reconstruction image block according to another part of the adjacent reconstruction image block. Double upsampling processing, in which part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
可選地,通過第一濾波器對當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對當前重建圖像區塊進行一次升取樣處理;相應的,方法還包括:若當前重建圖像區塊所處的當前圖像的所有圖像區塊均已完成重建,則通過第三濾波器,根據在所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,另一部分相鄰重建圖像區塊是在對所述當前重建圖像區塊進行第一次升取樣處理時未完成重建的圖像區塊;當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: the currently completed neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of a part of the reconstructed adjacent reconstructed image blocks are subjected to an up-sampling process of the current reconstructed image block through the first filter; correspondingly, the method further includes: if the current reconstructed image block is located All the image blocks of the image have been reconstructed, then pass the third filter to the current reconstructed image block according to another part of the adjacent reconstructed image block in the required adjacent reconstructed image block The part of the boundary is subjected to double upsampling, where another part of the adjacent reconstructed image block is an image block that has not been reconstructed when the first upsampling process is performed on the current reconstructed image block; the current Part of the boundary of the reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
可選的,其中第三濾波器可以是上述的第一濾波器,或者可以參照上述選擇第一濾波器的方式選擇第三濾波器,需要強調的是,該第三濾波器是針對當前重建圖像區塊的部分邊界進行升取樣處理的。Optionally, the third filter may be the above-mentioned first filter, or the third filter may be selected by referring to the above-mentioned way of selecting the first filter. It should be emphasized that the third filter is for the current reconstruction map Upsampling is performed on some boundaries of the block.
可選地,通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理之前,還包括:根據另一部分相鄰重建圖像區塊和當前重建圖像區塊的部分邊界判斷是否對部分邊界進行二次升取樣處理;若確定對第一重建圖像區塊的部分邊界進行二次升取樣處理,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, before the second upsampling process is performed on a part of the boundary of the current reconstructed image block according to another part of the neighboring reconstructed image block required by the third filter through the third filter, It also includes: judging whether to perform the sub-upsampling process on the partial boundary according to another part of the adjacent reconstructed image block and the partial boundary of the current reconstructed image block; For the sub-upsampling process, a third filter is used to perform a sub-upsampling process on part of the boundary of the current reconstructed image block according to another part of the adjacent reconstructed image block.
其中,上述方法與編碼端的對應方法相同,對應內容和效果在此不再贅述。Among them, the above method is the same as the corresponding method on the encoding side, and the corresponding content and effect will not be repeated here.
可選地,所述位元串流還包括:第二指示資訊;相應的,通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,包括:若第二指示資訊指示需要對當前重建圖像區塊進行二次升取樣處理,則通過第三濾波器,根據在所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理Optionally, the bit stream further includes: second indication information; correspondingly, through a third filter, another pair of adjacent reconstructed image block pairs according to the required adjacent reconstructed image block Performing sub-upsampling on part of the boundary of the current reconstructed image block, including: If the second indication information indicates that the current reconstructed image block needs to be up-sampled, then pass a third filter according to the required In the adjacent reconstructed image block of another part of the adjacent reconstructed image block, the partial boundary of the current reconstructed image block is up-sampled
通過這兩種可選方式可以判斷是否對當前重建圖像區塊進行二次升取樣處理。當確定不對當前重建圖像區塊進行二次升取樣處理時,可以降低解碼端的開銷,當確定需要對當前重建圖像區塊進行二次升取樣處理時,可以解決當前重建圖像區塊的邊界不連續問題。Through these two optional ways, it can be judged whether the current reconstructed image block is up-sampled. When it is determined that the current reconstructed image block is not to be upsampled, the overhead of the decoding end can be reduced. When it is determined that the current reconstructed image block needs to be upsampled, the current reconstructed image block can be resolved The problem of boundary discontinuity.
圖25為本發明再一實施例提供的一種圖像處理方法的流程圖,如圖25所示,該方法包括:FIG. 25 is a flowchart of an image processing method according to still another embodiment of the present invention. As shown in FIG. 25, the method includes:
步驟S2501:解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和當前待重建圖像區塊的編碼方式;Step S2501: Parse the bit stream to obtain the coding information of the current image block to be reconstructed and the coding method of the current image block to be reconstructed;
步驟S2502:根據編碼資訊生成當前待重建圖像區塊的重建訊號,並根據重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;Step S2502: generate a reconstruction signal of the current image block to be reconstructed according to the encoded information, and reconstruct the current image block to be reconstructed according to the reconstruction signal to obtain the current reconstruction image block;
步驟S2503:若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則從位元串流獲取第一濾波器的標識資訊,並通過標識資訊標識的第一濾波器對當前重建圖像區塊進行升取樣處理。Step S2503: If the current encoding method of the image block to be reconstructed is the down-sampling encoding method, then obtain the identification information of the first filter from the bit stream, and use the first filter identified by the identification information to the current reconstructed image The block is upsampled.
其中,該位元串流還包括:當前待重建圖像區塊的編碼方式、當前待重建圖像區塊項所述之當前圖像中已完成重建的重建圖像區塊的編碼方式。該編碼資訊用於生成待重建圖像區塊的重建訊號,比如:解碼端通過對編碼資訊進行亂度解碼,得到當前待重建圖像區塊的變換量化係數,然後對變換量化係數進行反量化和反變換,得到當前待重建圖像區塊的重建的殘差訊號。通過它的參考重建圖像區塊(參考重建圖像區塊的資訊屬於編碼資訊)對當前待重建圖像區塊進行預測,得到當前待重建圖像區塊的預測訊號,然後將預測訊號和上述重建的殘差訊號相加,得到當前待重建圖像區塊的重建訊號。該位元串流還包括現有技術涉及的更多的編碼資訊,本發明對此不做限制。Wherein, the bit stream also includes: the current encoding method of the image block to be reconstructed, and the encoding method of the reconstructed image block in the current image described in the current image block item to be reconstructed. The coding information is used to generate reconstruction signals of the image block to be reconstructed. For example, the decoding end obtains the transform and quantization coefficients of the current image block to be reconstructed by random decoding the encoded information, and then inversely quantizes the transform and quantization coefficients And inverse transform to obtain the reconstructed residual signal of the current image block to be reconstructed. Through its reference reconstruction image block (the information of the reference reconstruction image block belongs to the coding information), the current image block to be reconstructed is predicted to obtain the prediction signal of the current image block to be reconstructed, and then the prediction signal and The reconstructed residual signals are added to obtain the reconstruction signal of the current image block to be reconstructed. The bit stream also includes more coding information related to the prior art, which is not limited by the present invention.
最後,若當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則通過第一濾波器對當前重建圖像區塊進行升取樣處理。Finally, if the current reconstruction image block encoding method is a down-sampling encoding method, the current reconstruction image block is up-sampled through the first filter.
需要說明的是,解碼端和編碼端涉及的重建當前待重建圖像區塊方法類似,並且解碼端和編碼端涉及的對當前重建圖像區塊進行升取樣處理地方法也類似。在此不再贅述這兩個方法。It should be noted that the decoding end and the encoding end are involved in a method of reconstructing the current image block to be reconstructed, and the decoding end and the encoding end are involved in a method of upsampling the currently reconstructed image block. These two methods are not repeated here.
綜所上述,相對於現有技術中對整幅圖像中的重建圖像區塊都採用相同的濾波器,本發明針對每個重建圖像區塊都要選擇對應的濾波器,即本發明採用具有針對性的選擇方式來選擇濾波器,通過選擇的濾波器對重建圖像區塊進行升取樣處理,從而可以得到顯示效果更佳的重建圖像區塊。In summary, relative to the prior art, which uses the same filter for the reconstructed image blocks in the entire image, the present invention selects the corresponding filter for each reconstructed image block, that is, the present invention uses The filter is selected in a targeted selection manner, and the reconstructed image block is up-sampled by the selected filter, so that a reconstructed image block with better display effect can be obtained.
可選地,通過第一濾波器對當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對當前重建圖像區塊進行一次升取樣處理;相應的,所述方法還包括:若所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: the currently completed neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block A part of the reconstructed pixels of adjacent reconstructed image blocks are subjected to an up-sampling process on the current reconstructed image block through the first filter; correspondingly, the method further includes: If another part of the adjacent reconstruction image block in the block that has not been reconstructed currently has been reconstructed, the third filter is used to perform partial boundary reconstruction of the current reconstruction image block according to another part of the adjacent reconstruction image block. Double upsampling processing, in which part of the boundary of the current reconstructed image block is adjacent to another part of the adjacent reconstructed image block.
可選地,通過第一濾波器對當前重建圖像區塊進行升取樣處理,包括:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過第一濾波器對當前重建圖像區塊進行一次升取樣處理;相應的,方法還包括:若當前重建圖像區塊所處的當前圖像的所有圖像區塊均已完成重建,則通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,另一部分相鄰重建圖像區塊是在對所述當前重建圖像區塊進行第一次升取樣處理時未完成重建的圖像區塊;當前重建圖像區塊的部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, upsampling the current reconstructed image block through the first filter includes: the currently completed neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of a part of the reconstructed adjacent reconstructed image blocks are subjected to an up-sampling process of the current reconstructed image block through the first filter; correspondingly, the method further includes: if the current reconstructed image block is located After all image blocks of the image have been reconstructed, the third filter is used to determine the current reconstruction image block ’s Second boundary upsampling is performed on part of the boundary, where another part of the adjacent reconstructed image block is an image block that has not been reconstructed when the first upsampling process is performed on the current reconstructed image block; the current reconstruction Part of the boundary of the image block is adjacent to another part of the adjacent reconstructed image block.
可選地,通過第三濾波器,根據所述所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理之前,還包括:根據所述另一部分相鄰重建圖像區塊和所述當前重建圖像區塊的部分邊界判斷是否對所述部分邊界進行二次升取樣處理;若確定對第一重建圖像區塊的部分邊界進行二次升取樣處理,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, a third filter is used to secondarily raise the partial boundary of the current reconstructed image block according to another part of the required adjacent reconstructed image blocks Before the sampling process, it also includes: judging whether to perform a second upsampling process on the partial boundary according to the partial boundary of the another partial adjacent reconstructed image block and the current reconstructed image block; if it is determined that the first When the partial boundary of the reconstructed image block is subjected to sub-upsampling processing, a third filter is used to perform sub-upsampling processing on the partial boundary of the current reconstructed image block according to another partial adjacent reconstructed image block.
其中,上述方法與編碼端的對應方法相同,對應內容和效果在此不再贅述。Among them, the above method is the same as the corresponding method on the encoding side, and the corresponding content and effect will not be repeated here.
可選地,所述位元串流還包括:第二指示資訊;相應的,通過第三濾波器,根據所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,包括:若第二指示資訊指示需要對當前重建圖像區塊進行二次升取樣處理,則通過第三濾波器,根據在所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, the bit stream further includes: second indication information; correspondingly, through a third filter, another pair of adjacent reconstructed image block pairs according to the required adjacent reconstructed image block Performing sub-upsampling processing on a part of the boundary of the current reconstructed image block includes: if the second indication information indicates that it is necessary to perform up-sampling processing on the current reconstructed image block, passing a third filter according to Another part of the neighboring reconstructed image blocks in the required neighboring reconstructed image blocks performs sub-upsampling processing on part of the boundary of the current reconstructed image block.
通過這兩種可選方式可以判斷是否對當前重建圖像區塊進行二次升取樣處理。當確定不對當前重建圖像區塊進行二次升取樣處理時,可以降低解碼端的開銷,當確定需要對當前重建圖像區塊進行二次升取樣處理時,可以解決當前重建圖像區塊的邊界不連續問題。Through these two optional ways, it can be judged whether the current reconstructed image block is up-sampled. When it is determined that the current reconstructed image block is not to be upsampled, the overhead of the decoding end can be reduced. When it is determined that the current reconstructed image block needs to be upsampled, the current reconstructed image block can be resolved The problem of boundary discontinuity.
需要說明的是,本發明編碼端和解碼端還可以通過預先已協商的濾波器對當前重建圖像區塊進行升取樣處理。It should be noted that, the encoding end and the decoding end of the present invention may also perform up-sampling processing on the currently reconstructed image block through a previously negotiated filter.
圖26為本發明一實施例提供的一種圖像處理設備的結構示意圖,如圖26所示,該設備包括:生成模組2601,用於生成當前待編碼圖像區塊的重建訊號,並根據所述重建訊號重建所述當前待編碼圖像區塊,以得到當前重建圖像區塊;選擇模組2602,用於若所述當前重建圖像區塊的編碼方式為縮減取樣編碼方式,則從至少兩個候選濾波器中選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器;處理模組2603,用於通過所述第一濾波器對所述當前重建圖像區塊進行升取樣處理。FIG. 26 is a schematic structural diagram of an image processing device according to an embodiment of the present invention. As shown in FIG. 26, the device includes: a generating module 2601 for generating a reconstruction signal of an image block to be encoded, and according to The reconstruction signal reconstructs the current image block to be encoded to obtain the current reconstructed image block; the selection module 2602 is used if the encoding method of the current reconstructed image block is a down-sampling encoding method, then A first filter for upsampling the current reconstructed image block is selected from at least two candidate filters; a processing module 2603 is used to perform the current reconstructed image area through the first filter The block is upsampled.
可選地,選擇模組2602,具體用於根據所述當前重建圖像區塊的紋理特徵從至少兩個候選濾波器中選擇所述第一濾波器。Optionally, the selection module 2602 is specifically configured to select the first filter from at least two candidate filters according to the texture characteristics of the current reconstructed image block.
可選地,選擇模組2602具體用於:具體用於根據預設映射關係和所述當前重建圖像區塊的紋理特徵選擇所述第一濾波器,所述預設映射關係是包括所述當前重建圖像區塊的紋理特徵的預設紋理特徵和包括所述第一濾波器的所述至少兩個候選濾波器之間的映射關係。Optionally, the selection module 2602 is specifically used for: specifically used to select the first filter according to a preset mapping relationship and the texture feature of the current reconstructed image block, the preset mapping relationship includes the The mapping relationship between the preset texture feature of the texture feature of the current reconstructed image block and the at least two candidate filters including the first filter.
可選地,選擇模組2602,具體用於確定當前重建圖像區塊的至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊與所述當前重建圖像區塊的相似度,其中,所述至少兩個相鄰重建圖像區塊進行升取樣處理時對應至少兩個第二濾波器,所述至少兩個候選濾波器包括所述至少兩個第二濾波器;從所述至少兩個第二濾波器中選擇與所述當前重建圖像區塊相似度最高的相鄰重建圖像區塊對應的第二濾波器作為所述第一濾波器。Optionally, the selection module 2602 is specifically used to determine the relationship between each adjacent reconstructed image block and the current reconstructed image block in at least two adjacent reconstructed image blocks of the current reconstructed image block Similarity, wherein the at least two adjacent reconstructed image blocks correspond to at least two second filters when performing upsampling processing, and the at least two candidate filters include the at least two second filters; A second filter corresponding to the adjacent reconstructed image block with the highest similarity of the current reconstructed image block is selected as the first filter from the at least two second filters.
可選地,所述選擇模組2602具體用於:通過所述至少兩個候選濾波器分別對所述當前重建圖像區塊進行升取樣處理,得到所述至少兩個候選濾波器分別對應的升取樣圖像區塊;分別計算所述至少兩個候選濾波器對應的升取樣圖像區塊與所述當前重建圖像區塊對應的原始圖像區塊的誤差;將最小誤差對應的候選濾波器作為所述第一濾波器。Optionally, the selection module 2602 is specifically configured to separately upsample the current reconstructed image block through the at least two candidate filters to obtain corresponding Upsampling the image block; calculating the error of the upsampling image block corresponding to the at least two candidate filters and the original image block corresponding to the current reconstructed image block; the candidate corresponding to the smallest error The filter serves as the first filter.
可選地,所述處理模組2603具體用於:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過所述第一濾波器對所述當前重建圖像區塊進行一次升取樣處理;所述處理模組2603,還用於若所述所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則通過第三濾波器,根據所述另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,所述當前重建圖像區塊的所述部分邊界與所述另一部分相鄰重建圖像區塊鄰接。Optionally, the processing module 2603 is specifically configured to: according to the currently reconstructed part of the adjacent reconstructed image blocks of the adjacent reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of the block, up-sampling the current reconstructed image block through the first filter; the processing module 2603 is also used to determine if the desired adjacent reconstructed image block In the currently unreconstructed reconstruction of another part of the adjacent reconstructed image block in which reconstruction has been completed, a third filter is used to determine the portion of the current reconstructed image block according to the other part of the adjacent reconstructed image block The boundary is subjected to double upsampling processing, wherein the part of the boundary of the current reconstructed image block is adjacent to the other part of the adjacent reconstructed image block.
可選地,所述處理模組2603具體用於:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過所述第一濾波器對所述當前重建圖像區塊進行一次升取樣處理,其中,所述所需的相鄰重建圖像區塊為所述當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊;所述處理模組2603,若所述當前重建圖像區塊所處的當前圖像的所有圖像區塊均已完成重建,則通過第三濾波器,根據所述所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,所述另一部分相鄰重建圖像區塊是在對所述當前重建圖像區塊進行第一次升取樣處理時未完成重建的圖像區塊;所述當前重建圖像區塊的所述部分邊界與所述另一部分相鄰重建圖像區塊鄰接。Optionally, the processing module 2603 is specifically configured to: according to the currently reconstructed part of the adjacent reconstructed image blocks of the adjacent reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of the block, up-sampling the current reconstructed image block through the first filter, wherein the required adjacent reconstructed image block is the current reconstructed image block Neighboring reconstructed image blocks required for upsampling processing; the processing module 2603, if all image blocks of the current image where the current reconstructed image block is located have been reconstructed, then Through a third filter, perform a second upsampling process on a part of the boundary of the current reconstructed image block according to another part of the required adjacent reconstructed image block , The another part of the adjacent reconstructed image block is an image block that has not been reconstructed when the current reconstructed image block is first upsampled; the location of the current reconstructed image block Reconstruction graph of the boundary between the part and the other part Adjacent blocks.
可選地,所述第三濾波器為所述第一濾波器。Optionally, the third filter is the first filter.
可選地,還包括:判斷模組2604,用於根據所述另一部分相鄰重建圖像區塊和所述當前重建圖像區塊的部分邊界判斷是否對所述部分邊界進行二次升取樣處理;所述選擇模組2602,具體用於若所述判斷模組2604確定對所述第一重建圖像區塊的部分邊界進行二次升取樣處理,則通過所述第三濾波器,根據所述另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, the method further includes: a judgment module 2604, configured to judge whether to re-upsample the partial boundary according to the partial boundary of the another partial adjacent reconstructed image block and the partial boundary of the current reconstructed image block Processing; the selection module 2602 is specifically used to pass the third filter based on the third filter if the determination module 2604 determines to perform a second upsampling process on a part of the boundary of the first reconstructed image block The another part of the adjacent reconstructed image block performs sub-upsampling processing on a part of the boundary of the current reconstructed image block.
可選地,所述生成模組2601,還用於生成位元串流,所述位元串流包括:所述第一濾波器的標識資訊。Optionally, the generating module 2601 is further configured to generate a bit stream, the bit stream includes: identification information of the first filter.
可選地,所述生成模組2601,還用於生成位元串流,所述位元串流包括:第一指示資訊,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對所述當前重建圖像區塊進行升取樣處理時使用的濾波器。Optionally, the generating module 2601 is further configured to generate a bit stream, the bit stream includes: first indication information, and the first indication information is used to indicate how to select a pair from at least two candidate filters A filter used when the current reconstructed image block is up-sampled.
可選地,所述生成模組2601,還用於生成位元串流,所述位元串流包括:第二指示資訊,其中,所述第二指示資訊用於指示所述解碼端是否需要對所述當前重建圖像區塊進行二次升取樣處理。Optionally, the generating module 2601 is further configured to generate a bit stream, the bit stream includes: second indication information, wherein the second indication information is used to indicate whether the decoding end needs Performing a second upsampling process on the current reconstructed image block.
本發明提供的圖像處理設備可以執行上述圖14對應的圖像處理方法以及該方法的可選方式,其實現原理和技術效果類似,此處不再贅述。The image processing device provided by the present invention can execute the above-mentioned image processing method corresponding to FIG. 14 and optional methods of the method. The implementation principles and technical effects are similar, and are not repeated here.
圖27為本發明另一實施例提供的一種圖像處理設備的結構示意圖,如圖27所示,該設備包括:解析模組2701,用於解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和所述當前待重建圖像區塊的編碼方式;生成模組2702,用於根據所述當前待重建圖像區塊的編碼資訊生成當前待重建圖像區塊的重建訊號,並根據所述重建訊號重建當前待重建圖像區塊,得到當前重建圖像區塊;選擇模組2703,用於若當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則根據從位元串流獲取的第一指示資訊選擇用於對當前重建圖像區塊進行升取樣處理的第一濾波器,其中,第一指示資訊用於指示如何從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;處理模組2704,用於通過第一濾波器對當前重建圖像區塊進行升取樣處理。FIG. 27 is a schematic structural diagram of an image processing device according to another embodiment of the present invention. As shown in FIG. 27, the device includes: a parsing module 2701 for parsing a bit stream to obtain the current image to be reconstructed Encoding information of the block and the encoding method of the current image block to be reconstructed; a generating module 2702 is used to generate a reconstruction signal of the current image block to be reconstructed according to the encoding information of the current image block to be reconstructed , And reconstruct the current image block to be reconstructed according to the reconstruction signal to obtain the current reconstructed image block; the selection module 2703 is used to encode the current image block to be reconstructed as a down-sampling encoding method, according to The first indication information acquired from the bit stream selects a first filter for upsampling the current reconstructed image block, where the first indication information is used to indicate how to select a pair from at least two candidate filters The filter used when the current reconstructed image block is up-sampled; the processing module 2704 is used to up-sample the current reconstructed image block through the first filter.
可選地,第一指示資訊用於指示根據當前重建圖像區塊的紋理特徵從至少兩個候選濾波器選擇對當前重建圖像區塊進行升取樣處理時使用的濾波器;選擇模組2703具體用於:根據預設映射關係和所述當前重建圖像區塊的紋理特徵選擇所述第一濾波器,所述預設映射關係是包括所述當前重建圖像區塊的紋理特徵的預設紋理特徵和包括所述第一濾波器的所述至少兩個候選濾波器之間的映射關係。Optionally, the first indication information is used to instruct to select a filter used for upsampling the current reconstructed image block from at least two candidate filters according to the texture characteristics of the current reconstructed image block; the selection module 2703 Specifically, it is used to select the first filter according to a preset mapping relationship and the texture feature of the current reconstructed image block, and the preset mapping relationship is a pre-filter including the texture feature of the current reconstructed image block Assume a mapping relationship between texture features and the at least two candidate filters including the first filter.
可選地,第一指示資訊用於指示根據當前重建圖像區塊的相鄰重建圖像區塊從所述至少兩個候選濾波器選擇對所述當前重建圖像區塊進行升取樣處理時使用的濾波器;選擇模組2703具體用於:確定至少兩個相鄰重建圖像區塊中每個相鄰重建圖像區塊與所述當前重建圖像區塊的相似度,所述至少兩個相鄰重建圖像區塊進行升取樣處理時對應至少兩個第二濾波器,所述至少兩個候選濾波器包括所述至少兩個第二濾波器;從所述至少兩個第二濾波器中選擇相似度最高的相鄰重建圖像區塊對應的第二濾波器作為所述第一濾波器。Optionally, the first indication information is used to indicate when the currently reconstructed image block is up-sampled from the at least two candidate filters according to the adjacent reconstructed image block of the current reconstructed image block The filter used; the selection module 2703 is specifically used to determine the similarity between each adjacent reconstructed image block and the current reconstructed image block in at least two adjacent reconstructed image blocks, the at least When performing up-sampling processing on two adjacent reconstructed image blocks, corresponding to at least two second filters, the at least two candidate filters include the at least two second filters; from the at least two second filters Among the filters, a second filter corresponding to an adjacent reconstructed image block with the highest similarity is selected as the first filter.
可選地,所述處理模組2704具體用於:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過所述第一濾波器對所述當前重建圖像區塊進行一次升取樣處理、處理模組2704,還用於若所需的相鄰重建圖像區塊中的當前未完成重建的另一部分相鄰重建圖像區塊已完成重建,則通過第三濾波器,根據另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,當前重建圖像區塊的所述部分邊界與另一部分相鄰重建圖像區塊鄰接。Optionally, the processing module 2704 is specifically configured to: according to the currently reconstructed part of the neighboring reconstructed image blocks of the neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of the block are subjected to an up-sampling process of the current reconstructed image block through the first filter, and the processing module 2704 is also used to determine if the current If the reconstruction of another part of the adjacent reconstructed image block has been completed, then the third filter is used to upsample the part of the boundary of the current reconstructed image block according to another part of the adjacent reconstructed image block. Processing, wherein the partial boundary of the current reconstructed image block is adjacent to another partially adjacent reconstructed image block.
可選地,所述處理模組2704具體用於:根據當前重建圖像區塊進行升取樣處理時所需的相鄰重建圖像區塊中的當前已完成重建的一部分相鄰重建圖像區塊的畫素,通過所述第一濾波器對所述當前重建圖像區塊進行一次升取樣處理;所述處理模組2704,還用於若當前重建圖像區塊所處的所述當前圖像的所有圖像區塊均已完成重建,則通過第三濾波器,根據所述所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理,其中,所述另一部分相鄰重建圖像區塊是在對所述當前重建圖像區塊進行第一次升取樣處理時未完成重建的圖像區塊;所述當前重建圖像區塊的所述部分邊界與所述另一部分相鄰重建圖像區塊鄰接。Optionally, the processing module 2704 is specifically configured to: according to the currently reconstructed part of the neighboring reconstructed image blocks of the neighboring reconstructed image blocks required for upsampling according to the current reconstructed image block The pixels of the block, up-sampling the current reconstructed image block through the first filter; the processing module 2704 is also used to determine if the current reconstructed image block is located in the current All the image blocks of the image have been reconstructed, and then pass a third filter to the current reconstruction image according to another part of the required adjacent reconstructed image blocks A part of the boundary of the image block is subjected to the second upsampling process, wherein the other part of the adjacent reconstructed image block is a picture that has not been reconstructed when the first upsampling process is performed on the current reconstructed image block Image block; the part of the boundary of the current reconstructed image block is adjacent to the other part of the adjacent reconstructed image block.
可選地,所述第三濾波器為所述第一濾波器。Optionally, the third filter is the first filter.
可選地,還包括:判斷模組2705,用於根據所述另一部分相鄰重建圖像區塊和所述當前重建圖像區塊的部分邊界判斷是否對所述部分邊界進行二次升取樣處理;所述處理模組2704,具體用於若所述判斷模組2705確定對所述第一重建圖像區塊的部分邊界進行二次升取樣處理,則通過所述第三濾波器,根據所述另一部分相鄰重建圖像區塊對所述當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, it further includes: a judgment module 2705 for judging whether to up-sample the partial boundary based on the partial boundary of the another partial adjacent reconstructed image block and the current reconstructed image block Processing; the processing module 2704 is specifically used to pass the third filter according to the third filter if the judgment module 2705 determines to perform sub-upsampling processing on a part of the boundary of the first reconstructed image block The another part of the adjacent reconstructed image block performs sub-upsampling processing on a part of the boundary of the current reconstructed image block.
可選地,所述位元串流還包括:第二指示資訊;相應的,所述處理模組2704,具體用於若第二指示資訊指示需要對當前重建圖像區塊進行二次升取樣處理,則通過第三濾波器,根據在所需的相鄰重建圖像區塊中的另一部分相鄰重建圖像區塊對當前重建圖像區塊的部分邊界進行二次升取樣處理。Optionally, the bit stream further includes: second indication information; correspondingly, the processing module 2704 is specifically configured to, if the second indication information indicates that the current reconstructed image block needs to be up-sampled For the processing, a third filter is used to perform a second upsampling process on a part of the boundary of the current reconstructed image block according to another part of the neighboring reconstructed image block in the required adjacent reconstructed image block.
本發明提供的圖像處理設備可以執行上述圖24對應的圖像處理方法以及該方法的可選方式,其實現原理和技術效果類似,此處不再贅述。The image processing device provided by the present invention can execute the image processing method corresponding to the above-mentioned FIG. 24 and an optional manner of the method. The implementation principles and technical effects are similar, and will not be repeated here.
圖28為本發明再一實施例提供的一種圖像處理設備的結構示意圖,如圖28所示,該設備包括:解析模組2801,用於解析位元串流,以獲取當前待重建圖像區塊的編碼資訊和所述當前待重建圖像區塊的編碼方式;生成模組2802,用於根據所述編碼資訊生成當前待重建圖像區塊的重建訊號,並根據所述重建訊號重建所述當前待重建圖像區塊,得到當前重建圖像區塊;解析模組2801,還用於若所述當前待重建圖像區塊的編碼方式為縮減取樣編碼方式,則從所述位元串流獲取第一濾波器的標識資訊;處理模組2803,用於通過所述標識資訊標識的第一濾波器對所述當前重建圖像區塊進行升取樣處理。FIG. 28 is a schematic structural diagram of an image processing device according to still another embodiment of the present invention. As shown in FIG. 28, the device includes: a parsing module 2801 for parsing a bit stream to obtain the current image to be reconstructed The encoding information of the block and the encoding method of the current image block to be reconstructed; the generating module 2802 is used to generate a reconstruction signal of the current image block to be reconstructed according to the encoding information and reconstruct it according to the reconstruction signal The current image block to be reconstructed obtains the current image block to be reconstructed; the parsing module 2801 is also used to decode the bit from the bit if the encoding method of the current image block to be reconstructed is a down-sampling encoding method. The meta stream obtains the identification information of the first filter; the processing module 2803 is used for upsampling the current reconstructed image block through the first filter identified by the identification information.
本發明提供的圖像處理設備可以執行上述圖25對應的圖像處理方法以及該方法的可選方式,其實現原理和技術效果類似,此處不再贅述。The image processing device provided by the present invention can execute the image processing method corresponding to the above-mentioned FIG. 25 and optional methods of the method. The implementation principles and technical effects are similar, and details are not described here.
本發明提供一種圖像處理設備,該設備包括:處理器和用於儲存所述處理器的可執行指令的記憶體;其中該處理器可以執行圖14對應的圖像處理方法以及該方法的可選方式。其實現原理和技術效果類似,此處不再贅述。The present invention provides an image processing device including: a processor and a memory for storing executable instructions of the processor; wherein the processor can execute the image processing method corresponding to FIG. 14 and the method Election method. The implementation principle and technical effect are similar, and will not be repeated here.
本發明提供一種圖像處理設備,該設備包括:處理器和用於儲存所述處理器的可執行指令的記憶體;其中該處理器可以執行圖24對應的圖像處理方法以及該方法的可選方式。其實現原理和技術效果類似,此處不再贅述。The present invention provides an image processing device including: a processor and a memory for storing executable instructions of the processor; wherein the processor can execute the image processing method corresponding to FIG. 24 and the method Election method. The implementation principle and technical effect are similar, and will not be repeated here.
本發明提供一種圖像處理設備,該設備包括:處理器和用於儲存所述處理器的可執行指令的記憶體;其中該處理器可以執行圖25對應的圖像處理方法以及該方法的可選方式。其實現原理和技術效果類似,此處不再贅述。The present invention provides an image processing device including: a processor and a memory for storing executable instructions of the processor; wherein the processor can execute the image processing method corresponding to FIG. 25 and the method Election method. The implementation principle and technical effect are similar, and will not be repeated here.
圖29為本發明提供的一種圖像處理系統的結構示意圖,如圖29所示,該系統包括:上述解碼端的圖像處理設備2901,以及編碼端的圖像處理設備2902。FIG. 29 is a schematic structural diagram of an image processing system provided by the present invention. As shown in FIG. 29, the system includes the image processing device 2901 at the decoding end and the image processing device 2902 at the encoding end.
本發明提供的圖像處理系統包括的解碼端的圖像處理設備可以執行上述圖14對應的圖像處理方法以及該方法的可選方式,包括的編碼端的圖像處理設備可以執行上述圖24對應的圖像處理方法以及該方法的可選方式,其實現原理和技術效果類似,此處不再贅述。The image processing device included in the image processing system provided by the present invention can execute the image processing method corresponding to FIG. 14 described above and optional methods of the method, and the image processing device included in the encoding side can execute the image processing device corresponding to FIG. 24 described above. The image processing method and the optional method of the method have similar implementation principles and technical effects, and are not repeated here.
圖30為本發明提供的一種圖像處理系統的結構示意圖,如圖30所示,該系統包括:上述解碼端的圖像處理設備3001,以及編碼端的圖像處理設備3002。FIG. 30 is a schematic structural diagram of an image processing system provided by the present invention. As shown in FIG. 30, the system includes the image processing device 3001 at the decoding end and the image processing device 3002 at the encoding end.
本發明提供的圖像處理系統包括的解碼端的圖像處理設備可以執行上述圖14對應的圖像處理方法以及該方法的可選方式,包括的編碼端的圖像處理設備可以執行上述圖25對應的圖像處理方法以及該方法的可選方式,其實現原理和技術效果類似,此處不再贅述。The image processing device included in the image processing system provided by the present invention can execute the image processing method corresponding to the above-mentioned FIG. 14 and optional methods of the method, and the image processing device included on the encoding side can execute the above-mentioned image processing device corresponding to FIG. 25. The image processing method and the optional method of the method have similar implementation principles and technical effects, and are not repeated here.
雖然本發明已以實施例揭露如上,然其並非用以限定本發明,任何所屬技術領域中具有通常知識者,在不脫離本發明的精神和範圍內,當可作些許的更動與潤飾,故本發明的保護範圍當視後附的申請專利範圍所界定者為準。Although the present invention has been disclosed as above with examples, it is not intended to limit the present invention. Any person with ordinary knowledge in the technical field can make some changes and modifications without departing from the spirit and scope of the present invention. The scope of protection of the present invention shall be subject to the scope defined in the appended patent application.
S1401、S1402、S2401~ S2405、S2501~ S2503‧‧‧步驟S1401, S1402, S2401 ~ S2405, S2501 ~ S2503
2601‧‧‧生成模組2601‧‧‧Generation module
2602‧‧‧選擇模組2602‧‧‧Select module
2603‧‧‧處理模組2603‧‧‧Processing module
2604‧‧‧判斷模組2604‧‧‧judgment module
2701‧‧‧解析模組2701‧‧‧Analysis module
2702‧‧‧生成模組,2702‧‧‧Generation module,
2703‧‧‧選擇模組2703‧‧‧Select module
2704‧‧‧處理模組2704‧‧‧Processing module
2705‧‧‧判斷模組2705‧‧‧judgment module
2801‧‧‧解析模組2801‧‧‧Analysis module
2802‧‧‧生成模組2802‧‧‧Generation module
2803‧‧‧處理模組2803‧‧‧Processing module
2901、2902、3001、3002‧‧‧圖像處理設備2901, 2902, 3001, 3002‧‧‧ image processing equipment
圖1為本發明提供的數位影像的示意圖; 圖2為本發明提供的數位圖像的示意圖; 圖3為本發明提供的編碼端的編碼示意圖; 圖4為本發明提供的解碼端的解碼示意圖; 圖5為本發明提供的編碼端的編碼示意圖; 圖6為本發明一實施例提供的正在編碼的圖像的示意圖; 圖7為本發明一實施例提供的參考畫素範本的示意圖; 圖8A和圖8B為本發明一實施例提供的Planar模式的示意圖; 圖9為本發明一實施例提供的33種角度預測模式的具體方向示意圖; 圖10為本發明一實施例提供的圖像縮減取樣示意圖; 圖11為本發明一實施例提供的圖像升取樣示意圖; 圖12為本發明一實施例提供的圖像升取樣示意圖; 圖13為本發明一實施例提供的升取樣圖像示意圖; 圖14為本發明一實施例提供的一種圖像處理方法的流程圖; 圖15為本發明一實施例提供的4鄰域畫素的示意圖; 圖16為本發明一實施例提供的8鄰域畫素的示意圖; 圖17為本發明一實施例提供的圖像升取樣示意圖; 圖18為本發明另一實施例提供的圖像升取樣示意圖; 圖19為本發明一實施例提供的圖像升取樣示意圖; 圖20為本發明另一實施例提供的圖像升取樣示意圖; 圖21為本發明一實施例提供的當前重建圖像區塊和相鄰重建圖像區塊的示意圖; 圖22為本發明一實施例提供的當前重建圖像區塊的示意圖; 圖23為本發明一實施例提供的右邊界和右邊界的鄰接邊界的示意圖; 圖24為本發明另一實施例提供的一種圖像處理方法的流程圖; 圖25為本發明再一實施例提供的一種圖像處理方法的流程圖; 圖26為本發明一實施例提供的一種圖像處理設備的結構示意圖; 圖27為本發明另一實施例提供的一種圖像處理設備的結構示意圖; 圖28為本發明再一實施例提供的一種圖像處理設備的結構示意圖; 圖29為本發明提供的一種圖像處理系統的結構示意圖; 圖30為本發明提供的一種圖像處理系統的結構示意圖。1 is a schematic diagram of a digital image provided by the present invention; FIG. 2 is a schematic diagram of a digital image provided by the present invention; FIG. 3 is a schematic diagram of the encoding end provided by the present invention; FIG. 4 is a schematic diagram of decoding provided by the present invention; 5 is a schematic diagram of an encoding terminal provided by the present invention; FIG. 6 is a schematic diagram of an image being encoded according to an embodiment of the present invention; FIG. 7 is a schematic diagram of a reference pixel template provided by an embodiment of the present invention; FIG. 8A and FIG. 8B is a schematic diagram of the Planar mode provided by an embodiment of the present invention; FIG. 9 is a schematic diagram of specific directions of 33 angle prediction modes provided by an embodiment of the present invention; FIG. 10 is a schematic diagram of image downsampling provided by an embodiment of the present invention; 11 is a schematic diagram of image upsampling provided by an embodiment of the invention; FIG. 12 is a schematic diagram of image upsampling provided by an embodiment of the invention; FIG. 13 is a schematic diagram of upsampling image provided by an embodiment of the invention; FIG. 14 FIG. 15 is a flowchart of an image processing method according to an embodiment of the present invention; FIG. 15 is a schematic diagram of 4-neighbor pixels provided by an embodiment of the present invention; FIG. 16 A schematic diagram of 8 neighborhood pixels provided by an embodiment of the invention; FIG. 17 is a schematic diagram of image upsampling provided by an embodiment of the invention; FIG. 18 is a schematic diagram of image upsampling provided by another embodiment of the invention; FIG. 20 is a schematic diagram of image upsampling provided by an embodiment of the invention; FIG. 20 is a schematic diagram of image upsampling provided by another embodiment of the invention; FIG. 21 is a current reconstructed image block and neighbors provided by an embodiment of the invention A schematic diagram of a reconstructed image block; FIG. 22 is a schematic diagram of a current reconstructed image block provided by an embodiment of the invention; FIG. 23 is a schematic diagram of a right border and an adjacent border of a right border provided by an embodiment of the invention; FIG. 24 FIG. 25 is a flowchart of an image processing method according to another embodiment of the invention; FIG. 25 is a flowchart of an image processing method according to another embodiment of the invention; FIG. 26 is a diagram according to an embodiment of the invention Schematic diagram of an image processing device; FIG. 27 is a schematic diagram of an image processing device according to another embodiment of the invention; FIG. 28 is an image processing device according to another embodiment of the invention Structural diagram; schematic structural diagram of an image processing system according to the present invention is provided in FIG. 30; FIG. 29 a schematic structure of an image processing system provided by the present invention.
Claims (44)
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710571640.1A CN109257605B (en) | 2017-07-13 | 2017-07-13 | Image processing method, device and system |
CN201710571640.1 | 2017-07-13 | ||
??201710571640.1 | 2017-07-13 |
Publications (2)
Publication Number | Publication Date |
---|---|
TW201909643A true TW201909643A (en) | 2019-03-01 |
TWI681672B TWI681672B (en) | 2020-01-01 |
Family
ID=65001554
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW107116752A TWI681672B (en) | 2017-07-13 | 2018-05-17 | Method, apparatus and system for processing picture |
Country Status (3)
Country | Link |
---|---|
CN (1) | CN109257605B (en) |
TW (1) | TWI681672B (en) |
WO (1) | WO2019011046A1 (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20210125086A (en) * | 2019-02-22 | 2021-10-15 | 후아웨이 테크놀러지 컴퍼니 리미티드 | Method and apparatus for intra prediction using a linear model |
CN111369438B (en) * | 2020-02-28 | 2022-07-26 | 北京市商汤科技开发有限公司 | Image processing method and device, electronic equipment and storage medium |
WO2022237899A1 (en) * | 2021-05-14 | 2022-11-17 | Beijing Bytedance Network Technology Co., Ltd. | Method, device, and medium for video processing |
WO2023019567A1 (en) * | 2021-08-20 | 2023-02-23 | 深圳传音控股股份有限公司 | Image processing method, mobile terminal and storage medium |
CN113822815B (en) * | 2021-09-24 | 2024-02-06 | 广州光锥元信息科技有限公司 | Method and apparatus for high performance picture clutter removal using GPU rendering |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
BRPI0509563A (en) * | 2004-04-02 | 2007-09-25 | Thomson Licensing | scalable complexity video encoding |
US7961963B2 (en) * | 2005-03-18 | 2011-06-14 | Sharp Laboratories Of America, Inc. | Methods and systems for extended spatial scalability with picture-level adaptation |
CN101742321B (en) * | 2010-01-12 | 2011-07-27 | 浙江大学 | Layer decomposition-based Method and device for encoding and decoding video |
US10448032B2 (en) * | 2012-09-04 | 2019-10-15 | Qualcomm Incorporated | Signaling of down-sampling location information in scalable video coding |
CN103716622B (en) * | 2012-09-29 | 2017-02-08 | 华为技术有限公司 | Image processing method and device |
CN103916676B (en) * | 2012-12-31 | 2017-09-29 | 华为技术有限公司 | A kind of boundary intensity determines method, block-eliminating effect filtering method and device |
KR20140089487A (en) * | 2013-01-04 | 2014-07-15 | 삼성전자주식회사 | Method and apparatus for scalable video encoding using image upsampling based on phase-shift, method and apparatus for scalable video decoding using image upsampling based on phase-shift |
US10455249B2 (en) * | 2015-03-20 | 2019-10-22 | Qualcomm Incorporated | Downsampling process for linear model prediction mode |
-
2017
- 2017-07-13 CN CN201710571640.1A patent/CN109257605B/en active Active
-
2018
- 2018-05-04 WO PCT/CN2018/085537 patent/WO2019011046A1/en active Application Filing
- 2018-05-17 TW TW107116752A patent/TWI681672B/en active
Also Published As
Publication number | Publication date |
---|---|
WO2019011046A1 (en) | 2019-01-17 |
CN109257605B (en) | 2021-11-19 |
CN109257605A (en) | 2019-01-22 |
TWI681672B (en) | 2020-01-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
TWI681672B (en) | Method, apparatus and system for processing picture | |
TWI681670B (en) | Method, apparatus and system for processing picture | |
KR102185954B1 (en) | Apparatus and method for image coding and decoding | |
Wu et al. | Low bit-rate image compression via adaptive down-sampling and constrained least squares upconversion | |
TWI672941B (en) | Method, apparatus and system for processing picture | |
CN111586410B (en) | Video encoding method, decoding method and related devices thereof | |
KR101285810B1 (en) | A Super-Resolution method by Motion Estimation with Sub-pixel Accuracy using 6-Tap FIR Filter | |
KR101059649B1 (en) | Encoding device, method and decoding device using adaptive interpolation filter | |
WO2019190342A1 (en) | Image processing apparatus and method | |
Drynkin et al. | Video images compression and restoration methods based on optimal sampling | |
Moses et al. | A survey on adaptive image interpolation based on quantitative measures | |
Gupta et al. | An Efficient and Fast Method for Image Super Resolution by Block Based Image up Sampling | |
Doma | Comparison of different image interpolation algorithms | |
Patil et al. | Real time super resolution image reconstruction | |
VENKATESH et al. | Low Bit-Rate Image Compression Using CADU Compression |