WO2012079237A1 - 一种图像压缩中利用人眼视觉模型的小波系数量化方法 - Google Patents
一种图像压缩中利用人眼视觉模型的小波系数量化方法 Download PDFInfo
- Publication number
- WO2012079237A1 WO2012079237A1 PCT/CN2010/079875 CN2010079875W WO2012079237A1 WO 2012079237 A1 WO2012079237 A1 WO 2012079237A1 CN 2010079875 W CN2010079875 W CN 2010079875W WO 2012079237 A1 WO2012079237 A1 WO 2012079237A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- value
- component
- wavelet
- block
- frequency
- Prior art date
Links
- 230000000007 visual effect Effects 0.000 title claims abstract description 83
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000013139 quantization Methods 0.000 title claims abstract description 47
- 238000007906 compression Methods 0.000 title claims abstract description 46
- 230000006835 compression Effects 0.000 title claims abstract description 39
- 230000000873 masking effect Effects 0.000 claims abstract description 23
- 238000002474 experimental method Methods 0.000 claims abstract description 8
- 238000012360 testing method Methods 0.000 claims description 3
- CMRJPMODSSEAPL-FYQPLNBISA-N Trenbolone acetate Chemical compound C1CC2=CC(=O)CCC2=C2[C@@H]1[C@@H]1CC[C@H](OC(=O)C)[C@@]1(C)C=C2 CMRJPMODSSEAPL-FYQPLNBISA-N 0.000 claims 1
- 238000005192 partition Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 3
- 238000000638 solvent extraction Methods 0.000 abstract description 2
- 238000007619 statistical method Methods 0.000 abstract 1
- 230000008569 process Effects 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000003068 static effect Effects 0.000 description 4
- 230000006837 decompression Effects 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 230000000903 blocking effect Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000009982 effect on human Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/63—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets
-
- 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
- H04N19/124—Quantisation
-
- 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
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/154—Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion
-
- 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
- H04N19/18—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 a set of transform coefficients
Definitions
- the present invention relates to a wavelet coefficient quantization method for an image compression process, and more particularly to a wavelet coefficient quantization method for improving the subjective quality of a restored image by using human visual characteristics in an image compression process, and belongs to the field of static image compression technology.
- the above methods still have certain defects: When the compression factor is high, using the wavelet-based compression method, the restored image is prone to significant visual distortion. The reason is that the human eye as the ultimate recipient of the image has different sensitivity to different types of image distortion. Currently, the performance of the image compression method is judged by PNSR value or MSE, and this important factor is not considered.
- the classical image compression methods based on wavelet transform such as EBC0T and SPHIT
- the compression process is divided into four steps: preprocessing, wavelet transform, quantization, and encoding.
- preprocessing stage the chromaticity space conversion and the removal of the DC component are completed;
- the wavelet transform is used to reduce the spatial correlation of the image data, and the energy is concentrated in the low frequency portion to facilitate the compression;
- the quantization step is to reduce the entropy value of the image data. , is beneficial to improve the overall compression rate;
- entropy coding the quantized wavelet coefficients to obtain a compressed code stream.
- the quantization step is to cause the recovery map The only link like distortion.
- a good quantization strategy should be combined with human visual characteristics as much as possible, using different quantized values for different coefficients, discarding relatively insensitive and non-critical information of the human eye, and retaining sensitive and important information. Compressed while ensuring the quality of the restored image.
- the existing research shows that the main human visual characteristics related to images have brightness adaptability, masking effect and frequency characteristics, but in the current image compression method, human visual characteristics are rarely considered in the quantization stage, such as SPIHT and EZW are not considered. Human visual characteristics; even if a small part of the compression method takes into account the visual characteristics of the human eye, it is also very inadequate, mainly in the following aspects:
- the fixed visual weighting method calculates an image-independent image but is related to the line of sight (the distance between the human eye and the display).
- the sub-band weight table after wavelet transform, directly multiplies the coefficient by the corresponding weight, and then performs conventional coding.
- VIP vi sual progressive coding
- the model obtained by visual experiment in the airspace image does not establish contact with the wavelet coefficient, and can not directly reflect the sensitivity of different wavelet coefficients to the human visual system.
- DCQ dynamic contrastive Quantization
- the quantization value of each sub-band is dynamically adjusted according to the total contrast distortion.
- the visual distortion of the belt is equal, and the visual effect is optimal. Because the algorithm cannot directly establish the relationship between wavelet coefficients and visual quantized values, This leads to the need for iterative calculations, high complexity, and practicality.
- the technical problem to be solved by the present invention is to provide a wavelet coefficient quantization method using human visual mode in image compression. This method can effectively improve the subjective quality of the restored image, taking into account the accuracy of the visual model and the time and space complexity of the method.
- a method for quantizing a wavelet coefficient using a human visual model in image compression comprising the steps of:
- Step 1 Perform a wavelet transform on the compressed image, and divide all the obtained sub-bands into "X" blocks, and each block is denoted as S( ' ⁇ , 0, where ' represents the wavelet transform level of the sub-band where the block is located, ⁇ indicates the direction of the subband in which the block is located, indicating the position of the block in the subband;
- Step 2 Calculate the maximum value of the wavelet coefficients of the lowest frequency sub-band after the first-order wavelet transform, MAX k , the minimum value ⁇ / , and the average value of each block coefficient in the sub-band 0 (0, the average of the obtained block coefficients
- the value ⁇ (0 is linearly mapped to obtain its mapped gray value:
- the voltage at the pixel value of the CRT display is 0, and 6 is the value between the pixel value and the voltage value.
- the scale factor, c is the gamma value of the display, w is the Weber ratio, g is the energy gain factor of the wavelet transform at the first level; 7 and 2 are the upper and lower thresholds of the linear region in Weber's law, respectively;
- Step 3 The sub-bands of the sub-bands in the HL and LH directions after the second and third-order wavelet transforms are respectively obtained, and the masking component (_/, ⁇ ) of the visual model is calculated:
- Step 4 Calculate the joint component T G (j, d, i) of the two according to the luminance component and the mask component obtained above:
- T G (j, d, l) [T A (j, d, l) f [T M (j, d, l) where ", respectively represent the weight component of the luminance component and the mask component;
- Step 5 For the above The obtained joint component ⁇ ⁇ (_/ ⁇ , ) is quantized according to a preset level to obtain a gradation value ⁇ ⁇ ⁇ : ⁇ , and a value ⁇ after each joint component /) is quantized ( ⁇ , 0, encoding the level values ⁇ ⁇ ⁇ : ⁇ and the value ⁇ ( ⁇ 0 respectively;
- Step 6 selecting a plurality of test images having the same size as the image to be compressed, and obtaining the wavelet coefficients by using the above wavelet transform method, respectively
- the frequency component of the visual model is obtained for each frequency band as follows: The quantized value is adjusted, and the frequency band is quantized until the distortion of the restored image is just as perceived by the human eye, and the current quantized value is recorded as the image in the frequency
- the frequency JND value is averaged for the frequency JND value of the same frequency band in the above plurality of images to obtain the frequency component of the visual model / 0;
- Step 7 using the joint component ⁇ ⁇ (_/ ⁇ , ⁇ the quantized value ⁇ obtained above) ( ⁇ , 0 and visual
- the frequency component of the model, r s c / 0, calculates the visual quantization threshold ⁇ ( ⁇ 0:
- T(j, d, i) Q(j,d,i) - [T s (j,d,i)Y
- ⁇ represents the frequency component weight value
- Step 8 Quantify using the visual quantization threshold described above.
- the weight value of the luminance component when the compression ratio is less than 8 times, the weight value of the luminance component, "weight value masking component weights, a weight value of the frequency component of the right / to: 0.5, 0.5, 0.5; when the compression ratio is 8 ⁇ 20?
- the luminance component weight value when the compression ratio is 8 ⁇ 20?
- the masking component weight value ⁇ When is the luminance component weight value, the masking component weight value ⁇ , and the frequency component weight value? / For: 0. 5, 1. 0, 1. 0.
- the invention obtains a texture sensitivity model by using visual experiments, and calculates the quantization step size according to the texture features of the image block corresponding to different coefficient blocks, and finally quantizes the wavelet coefficients.
- the experimental results show that compared with the still image compression standard JPEG2000, the texture-based wavelet coefficient quantization method maintains the texture details of the original image and improves the subjective quality of the restored image.
- the image processing method provided by the present invention is particularly suitable for remote sensing image compression.
- Figure 1 is a block diagram of the prior art wavelet compression/decompression principle.
- FIG. 2 is a flow chart of image compression based on a wavelet domain visual quantization model according to the present invention.
- FIG. 3 is a schematic diagram of partitioning coefficients after wavelet transform in the present invention.
- Figure 4 is a schematic diagram of the calculation of frequency components in an iterative manner in the present invention.
- the wavelet coefficient quantization step in the image compression process is performed in the wavelet domain.
- JND Just Noticeable Differences
- each coefficient after the wavelet transform is componentized and inverse quantized, and only when the quantized value is less than a threshold D, the inverse transform
- a threshold D we define this threshold D as the JND value of the coefficient.
- the quantized value in each coefficient block can be adjusted according to the proportional relationship between the values in the quantization process, the restored image can achieve a good visual effect, which is visual quantization. Guiding ideology. However, each image has its own characteristics. If the JND is requantized according to the experimental method, it needs human judgment, and the process is cumbersome and has no practical value.
- the invention utilizes human visual characteristics, combined with the characteristics of wavelet transform, to establish a visual quantization model including a luminance component, a masking component and a frequency component in a wavelet domain, and then obtains a visual JND value as a visual characteristic of human visual characteristics through a visual quantization model.
- Quantitative value Considering three characteristics of the most dominant characteristics of human visual characteristics are luminance characteristics, frequency characteristics, and masking characteristics, the present invention decomposes visual quantized values into three components of a luminance component, a frequency component, and a masking component, and after obtaining the three components, Then, the weight values of the three components are obtained experimentally and combined to form the final visual model formula. This is the basic idea of the visual quantification model proposed by the present invention.
- the invention establishes a human eye visual quantization model for the coefficients of the image wavelet transform in units of blocks, and calculates a visual distortion threshold value of each coefficient block according to the model, and quantizes each block according to the threshold value, thereby ensuring that the quantized wavelet coefficients can be
- the visual information of the image is effectively preserved, and the visual effect of the decompressed image is optimized.
- the wavelet coefficient quantization method provided by the invention is particularly suitable for image compression with rich texture details such as remote sensing images.
- 2 is a wavelet coefficient quantization using the human visual model of the present invention Schematic diagram of the process.
- Step 1 Blocking the wavelet coefficients.
- the wavelet transform is performed on the compressed image, and all the obtained sub-bands are equally divided into "X" blocks, and each block is denoted as s, o, where 'is the wavelet transform level of the sub-band where the block is located, indicating the sub-band where the block is located
- the direction of the block indicates the position of the block in the sub-band; if the block in which the sub-band is split in this way is less than 2x 2, the sub-band does not need to be partitioned, and the entire sub-band is directly processed as a coefficient block.
- Figure 3 is a schematic diagram showing the position and size of the first block (black part) of each sub-band after a four-stage wavelet transform of an image of size 256x 256. After the fourth level transform, each sub-band is divided. Less than 2x 2, each sub-band becomes a single block (see the gray part in Figure 3).
- the quantized coefficients need to be stored in the compressed code stream and are easily dequantized during decompression, the visual quantized value is obtained in units of quantized coefficients, which leads to too many quantized values, and the code stream is too long and not feasible, so the comparison is performed.
- a good solution is to do block processing.
- the different sub-bands have different block sizes, which take into account the fact that the wavelet coefficients have the characteristics of parent-child correlation.
- the use of blocking can maintain the image detail feature and does not significantly increase the code traffic.
- the block of the upper frequency band and the corresponding block of the next level have parent-child correlation, when the luminance component and the mask component are obtained, only the recording is required. The result of block calculation at any level.
- Step 2 Calculate the luminance component.
- the greater the brightness of the image block, the stronger the tolerance to distortion, and the brightness distribution of the image block is related to the low-frequency coefficient block distribution in the wavelet transform process. Therefore, it can be obtained by the distribution of the low-frequency coefficient block in the wavelet transform process.
- the luminance distribution of the image block, and then the luminance components of each block are related to the low-frequency coefficient block distribution in the wavelet transform process.
- Step 3 Mask component calculation.
- the present invention establishes a masking model using the LH and HL band data of the 2- and 3-level transforms that are sensitive to the human eye, and uses the formula (3). ), calculate the masking component of the visual model (_/ ⁇ , ): jC(2, HL, ⁇ ) + C(3, HL, i) + ⁇ ]C(2, LH, i) + C(3, LH, i)
- C(2,H£,0 represents the maximum value of the absolute value of the coefficient in block S(2,H£,0, ( ⁇ 3 ' ⁇ ",, ⁇ shows the absolute value of the block S (3, H £, 0)
- the maximum value, C(2, LH, i) represents the maximum value of the absolute value of the coefficient in block B( 2 , LH, i), and C(3, LH, i) represents the coefficient in block B(3, LH, i) The maximum value of the absolute value.
- Step 4 Joint component calculation. According to the luminance component and the mask component obtained above, the joint component ⁇ ⁇ ( ⁇ ) of the two is calculated by the formula (4):
- T G (j, d, l) [T A (j, d, l) f [ M (j, d, l) f ( 4 )
- Step 5 Encode the joint component. It can be found from equations (2) and (3) that for any coefficient block, the luminance component r C/ ) and the masking component 0 are only related to each other, and are independent of _, d, so /) is only related, that is, All joint components actually have only "X" values corresponding to i. In order to facilitate the coding, it is first quantized into N levels, and N gradation values ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ are obtained , and the corresponding quantized values (l, d, after each block 7) (l, d, i) are quantized.
- ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ correspond to the index values 0 to N-1, respectively. Since these data need to be used in the decompression process, it is necessary to encode and store the gradation value ⁇ ⁇ in the compressed code stream, and the index value corresponding to the gradation value of each block quantization.
- Step 6 Count the frequency components.
- the frequency component calculation is independent of the image and wavelet coefficients to be encoded, but is related to the wavelet base used for the transform. Therefore, for a particular wavelet base, the frequency components can be obtained in a statistical manner. Referring to FIG. 4, in the present embodiment, a plurality of test images having different complexity are selected, and the wavelet coefficients are obtained by using the wavelet transform method described above, and the frequency components of the visual model are obtained for each frequency band as follows:
- Select image 1 to image n corresponding to the size of the image to be compressed, for a total of n images.
- Setting amount The initial value of the value is 2.0, and the frequency band is quantized and inverse quantized. Increasing or decreasing the quantized value, the frequency band is repeatedly quantized until the distortion of the restored image is just noticeable to the human eye, and the current quantized values Q1 to Qn are recorded as the frequency JND of the image 1 ⁇ n in the frequency band. value.
- the frequency JND values of the same frequency band in the above plurality of images are averaged to obtain a frequency component of the visual model).
- Step 7 Calculate the visual quantization threshold of each block according to the quantized value ⁇ ( ⁇ , 0 and the frequency component r s /, 0 of the visual model obtained by the joint component ⁇ ⁇ (_/ ⁇ , ⁇ ) obtained above.
- T(j,d,i) Q(j,d,i)-[T s (j,d,i)Y (5)
- ⁇ denotes a frequency component weight value.
- the luminance component weight value "change range is 0.4 to 0.6
- the mask component weight value ranges from 0.4 to 1.2
- the frequency component weight value ⁇ ranges from 0.4 to 1.2.
- the luminance component weight Value ", masking component weight value, frequency component weight value? / is preferably: 0.5, 0.5, 0.5; when the compression multiple is 8 to 20 times, the luminance component weight value", the masking component weight value, the frequency component weight value ? / is preferably: 0.5, 1.0, 1.0.
- Step 8 Quantify using the visual quantization threshold described above.
- the wavelet coefficient quantization method proposed by the present invention described above can be combined with any mainstream-based wavelet compression method, such as EZW, SPIHT, EBC0T, and the like.
- the experimental results show that after adding the quantization model of the present invention, under the same compression ratio, the texture information and small target of the image can be better preserved, and the subjective effect is better than the original method.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression Of Band Width Or Redundancy In Fax (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2010/079875 WO2012079237A1 (zh) | 2010-12-16 | 2010-12-16 | 一种图像压缩中利用人眼视觉模型的小波系数量化方法 |
CN201080044548.6A CN102934428B (zh) | 2010-12-16 | 2010-12-16 | 一种图像压缩中利用人眼视觉模型的小波系数量化方法 |
EP10858266.9A EP2654294A4 (en) | 2010-12-16 | 2010-12-16 | WAVELET COEFFICIENT QUANTIFICATION METHOD WITH A HUMAN VISUAL MODEL IN IMAGE COMPRESSION |
US13/500,070 US8655092B2 (en) | 2010-12-16 | 2010-12-16 | Wavelet coefficient quantization method using human visual model in image compression |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2010/079875 WO2012079237A1 (zh) | 2010-12-16 | 2010-12-16 | 一种图像压缩中利用人眼视觉模型的小波系数量化方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2012079237A1 true WO2012079237A1 (zh) | 2012-06-21 |
Family
ID=46243979
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2010/079875 WO2012079237A1 (zh) | 2010-12-16 | 2010-12-16 | 一种图像压缩中利用人眼视觉模型的小波系数量化方法 |
Country Status (4)
Country | Link |
---|---|
US (1) | US8655092B2 (zh) |
EP (1) | EP2654294A4 (zh) |
CN (1) | CN102934428B (zh) |
WO (1) | WO2012079237A1 (zh) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8811735B2 (en) * | 2012-07-14 | 2014-08-19 | Sigma Designs, Inc. | System and method for scalar quantization error reduction |
KR102071766B1 (ko) * | 2014-07-10 | 2020-03-02 | 인텔 코포레이션 | 효율적 텍스처 압축을 위한 방법 및 장치 |
CN104202594B (zh) * | 2014-07-25 | 2016-04-13 | 宁波大学 | 一种基于三维小波变换的视频质量评价方法 |
CN105491391A (zh) * | 2014-09-15 | 2016-04-13 | 联想(北京)有限公司 | 一种图像压缩方法及电子设备 |
CN107409213B (zh) | 2015-03-02 | 2020-10-30 | 杜比实验室特许公司 | 用于高动态范围图像的内容自适应感知量化器 |
CN108718409B (zh) * | 2018-06-11 | 2020-09-18 | 齐齐哈尔大学 | 一种遥感影像压缩方法 |
CN109191427B (zh) * | 2018-07-24 | 2020-09-22 | 中国传媒大学 | 基于双目抑制的dwt域亮度掩蔽模型构建方法及装置 |
CN111314704B (zh) * | 2018-12-12 | 2023-01-17 | 中国科学院深圳先进技术研究院 | 图像级jnd阈值的预测方法、装置、设备及存储介质 |
CN111651707B (zh) * | 2020-05-28 | 2023-04-25 | 广西大学 | 一种基于光学浅水区卫星遥感影像的潮位反演方法 |
CN112738515B (zh) * | 2020-12-28 | 2023-03-24 | 北京百度网讯科技有限公司 | 用于自适应量化的量化参数调整方法和装置 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2000074385A2 (en) * | 1999-05-27 | 2000-12-07 | University Of Maryland, College Park | 3d wavelet based video codec with human perceptual model |
CN1378384A (zh) * | 2001-03-23 | 2002-11-06 | 夏普公司 | 基于比特率预测和预测误差能量的自适应量化 |
CN1968419A (zh) * | 2005-11-16 | 2007-05-23 | 三星电子株式会社 | 利用人类视觉特性对图像进行编码及解码的方法和设备 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR100355375B1 (ko) * | 1995-11-01 | 2002-12-26 | 삼성전자 주식회사 | 영상부호화장치에있어서양자화간격결정방법및회로 |
US7006568B1 (en) * | 1999-05-27 | 2006-02-28 | University Of Maryland, College Park | 3D wavelet based video codec with human perceptual model |
US6792129B1 (en) * | 1999-08-25 | 2004-09-14 | Sharp Laboratories Of America | Digital watermarking in a perceptually uniform domain |
ES2305852T3 (es) * | 2003-10-10 | 2008-11-01 | Agency For Science, Technology And Research | Procedimiento de codificacion de una señal digital en un flujo binario escalable, procedimiento para la descodificacion de un flujo binario escalable. |
JP5001853B2 (ja) * | 2004-12-14 | 2012-08-15 | サムスン エレクトロニクス カンパニー リミテッド | 映像符号化及び復号化装置とその方法 |
-
2010
- 2010-12-16 CN CN201080044548.6A patent/CN102934428B/zh not_active Expired - Fee Related
- 2010-12-16 WO PCT/CN2010/079875 patent/WO2012079237A1/zh active Application Filing
- 2010-12-16 EP EP10858266.9A patent/EP2654294A4/en not_active Withdrawn
- 2010-12-16 US US13/500,070 patent/US8655092B2/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2000074385A2 (en) * | 1999-05-27 | 2000-12-07 | University Of Maryland, College Park | 3d wavelet based video codec with human perceptual model |
CN1378384A (zh) * | 2001-03-23 | 2002-11-06 | 夏普公司 | 基于比特率预测和预测误差能量的自适应量化 |
CN1968419A (zh) * | 2005-11-16 | 2007-05-23 | 三星电子株式会社 | 利用人类视觉特性对图像进行编码及解码的方法和设备 |
Non-Patent Citations (1)
Title |
---|
NGAN K N ET AL.: "VERY LOW BIT RATE VIDEO CODING USING 3D SUBBAND APPROACH", IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, vol. 4, no. 3, June 1994 (1994-06-01), pages 309 - 316, XP000460762 * |
Also Published As
Publication number | Publication date |
---|---|
CN102934428B (zh) | 2015-08-05 |
EP2654294A1 (en) | 2013-10-23 |
US8655092B2 (en) | 2014-02-18 |
US20130266237A1 (en) | 2013-10-10 |
EP2654294A4 (en) | 2015-11-11 |
CN102934428A (zh) | 2013-02-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2012079237A1 (zh) | 一种图像压缩中利用人眼视觉模型的小波系数量化方法 | |
JP6141295B2 (ja) | 知覚的に無損失のおよび知覚的に強調された画像圧縮システムならびに方法 | |
US7570817B2 (en) | Signal compression and decompression | |
JP2003521178A (ja) | 品質ベースのイメージ圧縮 | |
WO2010022002A1 (en) | Systems and methods for perceptually lossless video compression | |
JP2001298366A (ja) | 画像データ圧縮方法、及びデータ量子化方法及び装置 | |
CN108900838B (zh) | 一种基于hdr-vdp-2失真准则的率失真优化方法 | |
JPH09307904A (ja) | 映像信号符号化システム用量子化器 | |
CN103501438B (zh) | 一种基于主成分分析的内容自适应图像压缩方法 | |
EP1294175B1 (en) | Method and system for improving coding efficiency in image codecs | |
Kumar et al. | A review: DWT-DCT technique and arithmetic-Huffman coding based image compression | |
US8351723B2 (en) | Methods and apparatus for generating and using visual error weights | |
JP2005530423A (ja) | 画像鮮鋭度を符号化中に最適化する方法及びシステム | |
JP3105335B2 (ja) | 画像の直交変換符号化による圧縮・伸張方法 | |
CN103391437A (zh) | 一种高动态图像视觉无损压缩的方法及装置 | |
Zhang et al. | Visual distortion sensitivity modeling for spatially adaptive quantization in remote sensing image compression | |
US8478060B2 (en) | Method and device for compressing an image and storage medium carrying an image compressed by said method | |
Lamsrichan | Straightforward Color image compression using true-mean multi-level block truncation coding | |
Akbari et al. | Adaptive joint subband vector quantisation codec for handheld videophone applications | |
CN100566419C (zh) | 以无损的方式编码数字图像数据的设备和方法 | |
Chourasiya et al. | A Study of image compression based transmission algorithm Using SPIHT for low bit rate application | |
TWI404421B (zh) | 應用人類視覺恰可辨別差(jnd)模型之快速jpeg2000影像壓縮編碼方法及壓縮影像品質評估方法 | |
Liu | Perceptual jpeg2000 compliant coding for color images | |
Xu et al. | The performance analysis of wavelet in video coding system | |
Voukelatos et al. | A multiresolution adaptive VQ based still image codec with application to progressive image transmission |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201080044548.6 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13500070 Country of ref document: US |
|
REEP | Request for entry into the european phase |
Ref document number: 2010858266 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010858266 Country of ref document: EP |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 10858266 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |