WO2014103182A1 - 映像符号化装置、映像符号化方法および映像符号化プログラム - Google Patents
映像符号化装置、映像符号化方法および映像符号化プログラム Download PDFInfo
- Publication number
- WO2014103182A1 WO2014103182A1 PCT/JP2013/007080 JP2013007080W WO2014103182A1 WO 2014103182 A1 WO2014103182 A1 WO 2014103182A1 JP 2013007080 W JP2013007080 W JP 2013007080W WO 2014103182 A1 WO2014103182 A1 WO 2014103182A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- coefficient level
- value
- significant coefficient
- significant
- level
- Prior art date
Links
Images
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
- H04N19/124—Quantisation
- H04N19/126—Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
-
- 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
- H04N19/124—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
- H04N19/102—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
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- 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
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/147—Data rate or code amount at the encoder output according to rate distortion criteria
-
- 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
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/149—Data rate or code amount at the encoder output by estimating the code amount by means of a model, e.g. mathematical model or statistical model
-
- 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
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/15—Data rate or code amount at the encoder output by monitoring actual compressed data size at the memory before deciding storage at the transmission buffer
-
- 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
-
- 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
- H04N19/65—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using error resilience
-
- 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
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/91—Entropy coding, e.g. variable length coding [VLC] or arithmetic coding
Definitions
- the present invention is a quantization technique for quantizing orthogonal transform coefficients, and is suitably applied to, for example, a video encoding apparatus using HEVC.
- the video coding method based on the method described in Non-Patent Document 1 divides each frame of digitized video into coding tree units (CTU: Coding Tree Unit) and encodes each CTU in raster scan order. To do. Each CTU is divided into coding units (CU: Coding Unit) in a quad tree structure and encoded. Each CU is predicted by being divided into prediction units (PU: Prediction Unit). In addition, the prediction error of each CU is divided into transform units (TU: Transform Unit) in a quadtree structure, and is subjected to frequency conversion.
- CTU Coding Tree Unit
- CU is a coding unit for intra prediction / interframe prediction.
- intra prediction and inter-frame prediction will be described.
- Intra prediction is prediction generated from the reconstructed image of the encoding target frame.
- Non-patent document 1 defines 33 types of angle intra prediction shown in FIG.
- an intra prediction signal is generated by extrapolating the reconstructed pixels around the encoding target block in any of the 33 types of directions shown in FIG.
- a CU that uses intra prediction is referred to as an intra CU.
- Inter-frame prediction is prediction based on an image of a reconstructed frame (reference picture) having a display time different from that of an encoding target frame.
- inter-frame prediction is also referred to as inter prediction.
- FIG. 15 is an explanatory diagram illustrating an example of inter-frame prediction.
- the motion vector MV (mv x , mv y ) indicates the parallel movement amount of the reconstructed image block of the reference picture with respect to the encoding target block.
- Inter prediction generates an inter prediction signal based on a reconstructed image block of a reference picture (using pixel interpolation if necessary).
- a CU using inter prediction is referred to as an inter CU.
- a frame encoded only by the intra CU is called an I frame (or I picture).
- a frame encoded including not only an intra CU but also an inter CU is called a P frame (or P picture).
- a frame that is encoded including not only one reference picture for inter prediction of a block but also an inter CU that uses two reference pictures at the same time is called a B frame (or B picture).
- 16 includes a frequency converter 101, a quantizer 1020, an entropy encoder 103, an inverse frequency transformer / inverse quantizer 104, a buffer 105, a predictor 106, and an estimator 107.
- FIG. 17 is an explanatory diagram showing an example of CTU partitioning of frame t and a CU partitioning example of the eighth CTU (CTU8) of frame t when the spatial resolution of the frame is CIF (Common Intermediate Format) and the CTU size is 64.
- FIG. 18 is an explanatory diagram showing a quadtree structure corresponding to the CU partitioning example of CTU8.
- FIG. 19 is an explanatory diagram showing an example of TU partitioning of a CU.
- the upper part shows an example of intra prediction 2N ⁇ 2N PU CU TU partitioning.
- the CU is an intra prediction
- the root of the quad tree is arranged in the PU, and the prediction error is expressed by a quad tree structure.
- the bottom row shows an example of inter prediction 2N ⁇ N PU CU TU partitioning.
- the root of the quad tree is placed in the CU, and the prediction error is expressed by a quad tree structure.
- the estimator 107 determines a CU quadtree structure, a PU partition shape, and a TU quadtree structure that minimize the entropy coding cost for each CTU.
- the predictor 106 generates a prediction signal for the input image signal of the CU based on the CU quadtree structure and the PU partition shape determined by the estimator 107.
- the prediction signal is generated based on the above-described intra prediction or inter prediction.
- the frequency converter 101? Performs frequency conversion on the prediction error image obtained by subtracting the prediction signal from the input image signal based on the TU quadtree structure determined by the estimator 107 ?.
- the quantizer 1020 quantizes the frequency-transformed prediction error image (orthogonal transform coefficient).
- the quantized orthogonal transform coefficient is referred to as a coefficient level.
- a coefficient level having a value other than 0 is called a significant coefficient level.
- the quantizer 1020 includes a coefficient level calculation unit 1201 that receives an orthogonal transform coefficient Kij and a quantization parameter QP and outputs a coefficient level Lij.
- the entropy encoder 103 entropy-encodes cu_split_flag indicating a CTU quadtree structure, a prediction parameter, and a coefficient level.
- the inverse frequency transform / inverse quantizer 104 dequantizes the coefficient level. Further, the inverse frequency transform / inverse quantizer 104 performs inverse frequency transform on the inversely quantized orthogonal transform coefficient.
- the reconstructed prediction error image subjected to the inverse frequency conversion is supplied with a prediction signal and supplied to the buffer 105.
- the buffer 105 stores the reconstructed image.
- a general video encoding device Based on the above-described operation, a general video encoding device generates a bit stream.
- an orthogonal transformation coefficient Kij and a coefficient level Lij of 4 ⁇ 4 TU are defined as follows.
- Kij (0 ⁇ i, j ⁇ 3) is defined as the value of the orthogonal transformation coefficient at the horizontal position i and the vertical position j on the frequency axis.
- the coefficient level Lij is defined as the value of the coefficient level corresponding to the orthogonal transformation coefficient Kij.
- Kij and Lij are higher frequency components.
- the coefficient level calculation unit 1201 calculates the coefficient level Lij by dividing Kij by the quantization step Qs. Formally, the coefficient level Lij is expressed by equation (1).
- Lij Sign (Kij) / Floor (
- Sign (a) is a function that returns the sign of the input a
- Floor (a) is a function that returns the largest integer less than or equal to the input a
- f is a parameter that determines the quantization characteristics (0 ⁇ f ⁇ 0.5). is there.
- the value of f is 1/6 for inter prediction and 1/3 for intra prediction.
- N is the block size of TU.
- N 4.
- FIG. 21 shows an example in which Kij is quantized with Qs having a value of 4096 and f having a value of 1/3.
- the position information is information indicating the positions of all significant coefficient levels of the TU.
- the position information in HEVC includes information last_significant_x and last_significant_y indicating the horizontal position and vertical position of the significant coefficient level to be transmitted first, and the position subsequent to (last_significant_x, last_significant_y) ( 0, 0) and information siginificant_coeff_flag indicating the presence or absence of a significant coefficient level at each position. Therefore, the number of position information bits is the sum of the number of bits of last_significant_x, the number of bits of last_significant_y, and the number of bits of siginificant_coeff_flag determined based on the position of the significant coefficient level to be transmitted first.
- value information is information indicating the value of the significant coefficient level.
- the value information in HEVC is information coeff_abs_level_greater1_flag indicating whether or not the absolute value of the significant coefficient level is greater than 1, whether or not the absolute value of the significant coefficient level is greater than 2
- This information is composed of coeff_abs_level_remaining information.
- the number of value information bits is the sum of the number of bits of coeff_abs_level_greater1_flag, the number of bits of coeff_abs_level_greater2_flag, the number of bits of coeff_sign_flag, and the number of bits of coeff_abs_level_remaining.
- FIG. 22 shows the relationship between the position information and value information, and the coefficient level Lij.
- the items on the vertical axis with respect to the horizontal axis indicate information on each Lij in the 4 ⁇ 4 TU shown in FIG.
- siginificant_coeff_flag indicates the presence / absence of a significant coefficient level at each position from (3, 0) to (0, 0).
- siginificant_coeff_flag 1
- siginificant_coeff_flag 0.
- FIG. 23 shows the relationship between position information and value information and the number of bits of value information.
- the items on the vertical axis with respect to the horizontal axis indicate the number of position information bits and the number of value information bits of 4 ⁇ 4 TU shown in FIG.
- bin refers to one bit in the intermediate bit string before being converted to the bitstream output by the entropy encoder 103.
- the entropy encoder 103 ⁇ ⁇ ⁇ transmits bit 13bin of the position information of all significant coefficient levels in the TU and then the value information bits of each significant coefficient level. A total of 4 bins are transmitted.
- the position information is composed of last_significant_x, last_significant_y and siginificant_coeff_flag.
- siginificant_coeff_flag indicates the presence / absence of a significant coefficient level at each position of 9 from the position (2,) 1) to (0, ⁇ 0) following the position of the significant coefficient to be transmitted first, and is 9 bin.
- the value information is composed of coeff_abs_level_greater1_flag, coeff_coeff_abs_level_greater2_flag, coeff_sign_flag, and coeff_abs_level_remaining.
- coeff_abs_level_greater1_flag compared L 30 and L 01, are 2bin indicate whether 1 or greater.
- coeff_coeff_abs_level_greater2_flag is 0bin because there is no coefficient level with an absolute value of a significant coefficient level greater than 2.
- coeff_sign_flag indicates the sign of L 30 and L 01 and is 2 bins.
- coeff_abs_level_remaining is 0bin because there is no coefficient level with an absolute value of the significant coefficient level greater than 2.
- HEVC High efficiency video coding
- JCT-VC Joint Collaborative Team on Video Coding
- the entropy encoder using HEVC entropy encodes the position information of all significant coefficient levels in the TU in units of TUs, and then entropy codes the value information of each significant coefficient level. Therefore, there is a significant coefficient level (hereinafter referred to as a high cost coefficient level) that is expensive in transmission when the number of bits of position information and the number of bits of value information satisfy a predetermined condition. For example, when the number of bits of position information is larger than the number of bits of value information, the coefficient level becomes a high cost coefficient level. Further, when the value of the significant coefficient level is small, the coefficient level tends to become a high cost coefficient level. If the high cost coefficient level is transmitted, the compression efficiency decreases.
- a significant coefficient level hereinafter referred to as a high cost coefficient level
- An object of the present invention is to provide a video encoding device that prevents transmission at a high cost coefficient level and prevents a reduction in compression efficiency.
- a video encoding apparatus includes an orthogonal transform unit that orthogonally transforms an image block to calculate an orthogonal transform coefficient, a quantization unit that quantizes the orthogonal transform coefficient to calculate a coefficient level, and all of the coefficient levels.
- An entropy encoding unit that entropy-encodes each significant coefficient level value information and outputs a bitstream after entropy-encoding the position information of the significant coefficient level, and the quantization unit ,
- Position information bit number calculating means for calculating the number of position information bits determined based on the position of the significant coefficient level first in the transmission order included in the image block, and calculating the value information bit number of the significant coefficient level
- the video coding method orthogonally transforms an image block to calculate orthogonal transform coefficients, quantize the orthogonal transform coefficients to calculate coefficient levels, and entropy the position information of all significant coefficient levels among the coefficient levels.
- a video encoding method in which, after encoding, the value information of each significant coefficient level is entropy encoded and a bit stream is output, and when calculating the coefficient level, the first significant in the transmission order included in the image block The number of position information bits determined based on the position of the correct coefficient level is calculated, the value information bit number of the significant coefficient level is calculated, and the position information bit number and the value information bit number The coefficient level is set to 0.
- the video encoding program allows a computer to orthogonally transform an image block to calculate orthogonal transform coefficients, to quantize the orthogonal transform coefficients to calculate coefficient levels, and to calculate all significant values among the coefficient levels.
- After entropy coding the coefficient level position information, entropy coding of each significant coefficient level value information and executing a process of outputting a bit stream, and calculating the coefficient level, transmission included in the image block The process of calculating the number of position information bits determined based on the position of the first significant coefficient level in order, the process of calculating the number of value information bits of the significant coefficient level, the number of position information bits and the number of value information bits And a process of setting a significant coefficient level satisfying a predetermined condition to 0.
- FIG. 1 is a block diagram showing a first embodiment of a video encoding device according to the present invention.
- FIG. It is a block diagram which shows the structure of the adaptive quantizer in 1st Embodiment. It is a flowchart which shows operation
- FIG. 1 It is a block diagram which shows the structure of the adaptive quantizer in 5th Embodiment. It is a flowchart which shows operation
- FIG. 6 is an explanatory diagram showing an example of CTU partitioning of frame t and a CU partitioning example of CTU8 of frame t. It is explanatory drawing which shows the quadtree structure corresponding to the CU division
- FIG. FIG. 1 is a block diagram showing a first embodiment of a video encoding apparatus according to the present invention.
- the configuration of a video encoding apparatus according to a first embodiment that outputs a bit stream using each frame of a digitized video as an input image will be described.
- the video encoding apparatus includes a frequency converter 101, an adaptive quantizer 1021, an entropy encoder 103, an inverse frequency transformer / inverse quantizer 104, a buffer 105, a predictor 106, And an estimator 107.
- an adaptive quantizer 1021 is provided instead of the quantizer 1020.
- Other blocks in the video encoding device shown in FIG. 1B are the same as the blocks in the video encoding device shown in FIG.
- the entropy encoder 103 performs entropy encoding on the position information of all significant coefficient levels in the TU in units of TU, and then entropy encodes the value information of each significant coefficient level. Therefore, only the adaptive quantizer 1021 will be described below.
- FIG. 2 is a block diagram showing the configuration of adaptive quantizer 1021.
- the adaptive quantizer shown in FIG. 2 includes a value information bit number calculation unit (R value calculation unit) 1202 and a position information bit number calculation unit (R locate ) in addition to the coefficient level calculation unit 1201 of the quantizer shown in FIG. Calculation unit) 1203 and high cost coefficient level detection / removal unit 1204.
- R value calculation unit value information bit number calculation unit
- R locate position information bit number calculation unit
- the coefficient level calculation unit 1201 inputs the orthogonal transform coefficient Kij and the quantization parameter QP, and outputs a temporary coefficient level L′ ij.
- the R value calculation unit 1202 receives the temporary coefficient level L′ ij and outputs R value that is the number of value information bits.
- R value is the number of bits of information indicating the value of L′ ij. Specifically, R value is the number of bits of information coeff_abs_level_greater1_flag indicating whether or not the absolute value of the significant coefficient level is greater than 1, and the number of bits of information coeff_coeff_abs_level_greater2_flag that indicates whether or not the absolute value of the significant coefficient level is greater than 2.
- the R locate calculation unit 1203 receives the orthogonal transform coefficient position (i, j) and outputs R locate which is the number of position information bits.
- R locate is the number of bits of information indicating the positions on the frequency axis of all significant coefficient levels of TU. Specifically, R locate is the number of bits of last_significant_x and last_significant_y indicating the horizontal position and vertical position of the first significant coefficient level transmitted, and the position following (last_significant_x, last_significant_y) (0, 0) This is the sum of the number of bits of information siginificant_coeff_flag indicating the presence / absence of a significant coefficient level at positions up to.
- the high cost coefficient level detection / removal unit 1204 receives L′ ij, R value, and R locate and outputs the coefficient level Lij. For example, when R locate is greater than 0 and R locate is greater than R value , the high cost coefficient level detection / removal unit 1204 detects L'ij as the high cost coefficient level and outputs a value of 0 as Lij To do. Otherwise, the temporary coefficient level L′ ij is output as Lij. That is, the high cost coefficient level detection / removal unit 1204 detects and removes the high cost coefficient level.
- the adaptive quantizer 1021 performs transmission order ((3, 3) ⁇ (3, 2) ⁇ (2, for each orthogonal transform coefficient Kij (0 ⁇ i, j ⁇ 3) in each TU. 3) ⁇ (3, 1) ⁇ (2, 2) ⁇ (1, 3) ⁇ (3, 0) ⁇ (2, 1) ⁇ (1, 2) ⁇ (0, 3) ⁇ (2, 0) ⁇ (1, 1) ⁇ (0, 2) ⁇ (1, 0) ⁇ (0, 1) ⁇ (0, 0) in this order)
- step S101 the coefficient level calculation unit 1201 calculates a temporary coefficient level L′ ij corresponding to the orthogonal transformation coefficient Kij using, for example, the above equation (1).
- step S102 If it is determined in step S102 that the value of L'ij is not 0, the process proceeds to step S103. When the value of L′ ij is 0, the process proceeds to step S106.
- step S103 the R value calculation unit 1202 calculates the bit number R value of the value information of L′ ij.
- step S104 the R locate calculation unit 1203 determines whether a non-zero Lij has already been output in the TU. If a non-zero Lij is not yet output in the TU, the process proceeds to step S105. If non-zero Lij is already output in the TU, the value of R locate is set to 0, and the process proceeds to step S106.
- step S105 the R locate calculation unit 1203 calculates the number of bits R locate of the position information determined based on the position of L′ ij.
- the adaptive quantizer 1021 according to the present embodiment, the high cost factor level detection / removal section 1204, if greater than R The locate is 0, and R The locate is larger than the R value, L '
- the video encoding apparatus can prevent compression efficiency from being lowered by preventing transmission at a high cost coefficient level.
- Embodiment 2 FIG. In the first embodiment described above, the high cost coefficient level may not be detected accurately. Therefore, in order to detect the high cost coefficient level more accurately, the adaptive quantizer 1022 having the configuration shown in FIG. 4 is used in the second embodiment. Note that the video encoding apparatus of the second embodiment is basically configured as shown in FIG. 1, but an adaptive quantizer 1022 is provided instead of the adaptive quantizer 1021.
- the adaptive quantizer 1022 shown in FIG. 4 includes a coefficient level calculation unit 1201, a value information bit number calculation unit (R value calculation unit) 1202, a position information bit number calculation unit (R locate calculation unit) 1203, and a high cost coefficient level detection. / Removal unit 1204, and square error reduction amount calculation unit (D calculation unit) 1205.
- R value calculation unit value information bit number calculation unit
- R locate calculation unit position information bit number calculation unit
- D calculation unit square error reduction amount calculation unit
- the coefficient level calculation unit 1201 inputs the orthogonal transform coefficient Kij and the quantization parameter QP, and outputs a temporary coefficient level L′ ij.
- the D calculation unit 1205 receives the quantization parameter QP and L′ ij and outputs a square error reduction amount D.
- the square error reduction amount D is a square error reduction amount with respect to Kij.
- the D calculation unit 1205 uses the product of the square value of the quantization step Qs and the square value of L′ ij as shown in the following equation (3) to calculate the square error. Calculate the reduction amount D.
- the R value calculation unit 1202 receives L′ ij and outputs R value that is the number of value information bits.
- the R locate calculation unit 1203 receives the orthogonal transform coefficient position (i, j) and outputs R locate which is the number of position information bits.
- the high cost coefficient level detection / removal unit 1204 receives L′ ij, R value , R locate , and D and outputs the coefficient level Lij.
- the high cost coefficient level detection / removal unit 1204 uses the slope ⁇ of D with respect to the sum of R locate and R value .
- ⁇ is the gradient of the relationship between the square error reduction amount and the transmission code amount. ⁇ depends on the quantization parameter QP, and becomes smaller as the quantization parameter becomes smaller (as the quantization step size becomes smaller), and becomes larger as the quantization parameter becomes larger (as the quantization step size becomes larger). In the case of uniform quantization, ⁇ is expressed by the following equation (4) using Qs.
- the high cost coefficient level detection / removal unit 1204 increases L'ij when the absolute value of L'ij is greater than 0 and the product of R locate and R value and ⁇ is greater than or equal to D.
- the cost coefficient level is determined and a value of 0 is output as Lij. If L'ij is not at a high cost coefficient level, L'ij is output as Lij.
- the adaptive quantizer 1022 performs transmission order ((3, 3) ⁇ (3, 2) ⁇ (2, for each orthogonal transform coefficient Kij (0 ⁇ i, j ⁇ 3) in each TU. 3) ⁇ (3, 1) ⁇ (2, 2) ⁇ (1, 3) ⁇ (3, 0) ⁇ (2, 1) ⁇ (1, 2) ⁇ (0, 3) ⁇ (2, 0) ⁇ (1, 1) ⁇ (0, 2) ⁇ (1, 0) ⁇ (0, 1) ⁇ (0, 0) in this order)
- step S201 the coefficient level calculation unit 1201 calculates a temporary coefficient level L′ ij corresponding to the orthogonal transform coefficient Kij using, for example, the above equation (1).
- step S202 If it is determined in step S202 that the value of L'ij is not 0, the process proceeds to step S203. If the value of L′ ij is 0, the process proceeds to step S207.
- step S203 the D calculation unit 1205 calculates D that is a square error reduction amount for the orthogonal transform coefficient.
- step S204 the R value calculation unit 1202 calculates the bit number R value of the value information of L′ ij.
- step S205 the R locate calculation unit 1203 determines whether non-zero Lij has already been output in the TU. If non-zero Lij is not yet output in the TU, the process proceeds to step S206. If non-zero Lij has already been output in the TU, R locate is not calculated (the value of R locate is set to 0), and the process proceeds to step S207.
- step S206 the R locate calculation unit 1203 calculates the number of bits R locate of the position information determined based on the position of L′ ij.
- the adaptive quantizer 1022 in the present embodiment uses the square error reduction amount for Kij calculated by the square error reduction amount calculation unit (D calculation unit) 1205, whereby the adaptive quantizer in the first embodiment. Higher cost coefficient level can be detected more accurately than 1021. Therefore, in the video encoding device of the present embodiment, the effect of not reducing the compression efficiency is further increased by preventing transmission at a high cost coefficient level.
- Embodiment 3 FIG.
- the adaptive quantizer 1023 having the configuration shown in FIG. 6 is used in the third embodiment. Note that the video encoding apparatus of the third embodiment is basically configured as shown in FIG. 1, but an adaptive quantizer 1023 is provided instead of the adaptive quantizer 1021.
- the adaptive quantizer 1023 shown in FIG. 6 includes a coefficient level calculation unit 1201, a value information bit number calculation unit (R value calculation unit) 1202, a position information bit number calculation unit (R locate calculation unit) 1203, and a high cost coefficient level detection. / Removal unit 1204, and square error reduction amount calculation unit (D calculation unit) 1205.
- the coefficient level calculation unit 1201 receives the orthogonal transform coefficient Kij and the quantization parameter QP, calculates the temporary coefficient level L′ ij using, for example, the above equation (1), and outputs the temporary coefficient level L′ ij. .
- the D calculation unit 1205 receives the quantization parameters QP, L′ ij, and Kij, and outputs a square error reduction amount D.
- the square error reduction amount D is a square error reduction amount with respect to Kij.
- the D calculation unit 1205 calculates D using Kij, Qs, and L′ ij as shown in the following equation (5).
- the R value calculation unit 1202 receives L′ ij and outputs R value that is the number of value information bits.
- the R locate calculation unit 1203 inputs the orthogonal transform coefficient position (i, j) and outputs R locate which is the number of position information bits.
- the high cost coefficient level detection / removal unit 1204 receives L′ ij, R value , R locate , and D and outputs the coefficient level Lij.
- the high cost coefficient level detection / removal unit 1204 increases L'ij when the absolute value of L'ij is greater than 0 and the product of R locate and R value and ⁇ is greater than or equal to D. Judge as the cost coefficient level and output a value of 0 as Lij. If L'ij is not at a high cost coefficient level, L'ij is output as Lij.
- the D calculation unit 1205 calculates D using the quantization parameters QP and L′ ij, whereas in this embodiment, the D calculation unit 1205 includes Kij, Qs, and D is calculated using L′ ij, but the other processing is the same as in the second embodiment.
- the adaptive quantizer 1023 in the present embodiment uses the square error reduction amount for Kij calculated by the D calculator 1205 that inputs the quantization parameters QP, L'ij, and Kij ⁇ ⁇ , so that the second embodiment is more effective. Can accurately detect the high cost coefficient level. Therefore, in the video encoding apparatus of the present embodiment, the effect of not reducing the compression efficiency is further increased by preventing transmission at a high cost coefficient level.
- FIG. 7 is a block diagram illustrating a configuration of the adaptive quantizer 1024 in the video encoding device according to the fourth embodiment.
- Adaptive quantizer 1024 includes the maximum number of bits of coeff_sign_flag determined from the position of the significant coefficient level to be transmitted first in the number of position information bits. Note that the video encoding apparatus of the fourth embodiment is basically configured as shown in FIG. 1, but an adaptive quantizer 1024 is provided instead of the adaptive quantizer 1021.
- adaptive quantizer 1024 includes a coefficient level calculation unit 1201, an absolute value information bit number calculation unit (R abs calculation unit) 1206, a position / code number information bit number calculation unit (R locate_sign calculation unit) 1207, and A high cost coefficient level detection / removal unit 1204 is provided.
- R abs calculation unit absolute value information bit number calculation unit
- R locate_sign calculation unit position / code number information bit number calculation unit
- the coefficient level calculation unit 1201 receives the orthogonal transform coefficient Kij and the quantization parameter QP, and outputs a temporary coefficient level L′ ij.
- the R abs calculation unit 1206 receives L′ ij and outputs R abs which is the number of absolute value information bits.
- R abs is the number of bits of information indicating the absolute value of L′ ij. Specifically, R abs is the number of bits of information coeff_abs_level_greater1_flag indicating whether the absolute value of the significant coefficient level is greater than 1, the number of bits of information coeff_coeff_abs_level_greater2_flag indicating whether the absolute value of the significant coefficient level is greater than 2 And the sum of the number of bits of information coeff_abs_level_remaining calculated based on the description in 9.2.2.8 of Non-Patent Document 1 and indicating the absolute value of the value obtained by subtracting coeff_abs_level_greater1_flag and coeff_abs_level_greater2_flag from the absolute value of the significant coefficient level.
- the R locate_sign calculation unit 1207 receives the orthogonal transform coefficient position (i, j) and outputs R locate_sign which is the number of bits of position / code number information.
- R locate_sign is the number of bits of information indicating the positions on the frequency axis of all significant coefficient levels of TU.
- R locate_sign is the number of bits of last_significant_x and last_significant_y information indicating the horizontal position and vertical position of the first significant coefficient level to be transmitted, the position from (0, 0) to the position after (last_significant_x, last_significant_y) Is the sum of the number of bits of information significant_coeff_flag indicating the presence / absence of a significant coefficient level and the maximum number of bits of information coeff_sign_flag indicating the sign of the significant coefficient level.
- the high cost coefficient level detection / removal unit 1204 receives L′ ij, R abs , and R locate_sign and outputs the coefficient level Lij.
- R Locate_sign is greater than 0, and R Locate_sign is greater than R abs, detects L'ij as a high cost factor level, and outputs a value of 0 as Lij. Otherwise, the temporary coefficient level L′ ij is output as Lij.
- the adaptive quantizer 1024 performs transmission order ((3, 3) ⁇ (3, 2) ⁇ (2, for each orthogonal transform coefficient Kij (0 ⁇ i, j 3) in each TU. 3) ⁇ (3, 1) ⁇ (2, 2) ⁇ (1, 3) ⁇ (3, 0) ⁇ (2, 1) ⁇ (1, 2) ⁇ (0, 3) ⁇ (2, 0) ⁇ (1, 1) ⁇ (0, 2) ⁇ (1, 0) ⁇ (0, 1) ⁇ (0, 0) in this order)
- step S301 the coefficient level calculation unit 1201 calculates a temporary coefficient level L′ ij corresponding to the orthogonal transform coefficient Kij using, for example, the above equation (1).
- step S302 If it is determined in step S302 that the value of L'ij is not 0, the process proceeds to step S303. If the value of L′ ij is 0, the process proceeds to step S306.
- step S303 the R abs calculation unit 1206 calculates the number of bits R abs of the absolute value information of L′ ij.
- step S304 the R locate_sign calculation unit 1207 determines whether or not a non-zero Lij has already been output in the TU. If a non-zero Lij is not yet output in the TU, the process proceeds to step S305. If non-zero Lij is already output in the TU, the value of R locate_sign is set to 0, and the process proceeds to step S306.
- step S305 the R locate_sign calculation unit 1207 calculates the bit number R locate_sign of the position / code number information determined based on the position of L′ ij.
- the high cost factor level detection / removal section 1204, R Locate_sign is greater than 0, and when R Locate_sign is greater than R abs, detect L'ij as a high cost factor levels However, by outputting a value of 0 as Lij, the high cost coefficient level is not transmitted. Therefore, the video encoding apparatus according to the present embodiment can prevent compression efficiency from being lowered by preventing transmission at a high cost coefficient level.
- FIG. FIG. 9 is a block diagram showing the configuration of the adaptive quantizer 1025 in the video encoding device of the fifth embodiment.
- the adaptive quantizer 1025 includes the maximum number of bits of coeff_sign_flag determined from the position of the significant coefficient level to be transmitted first in the number of position information bits.
- the adaptive quantizer 1025 uses the square error reduction amount for Kij calculated by the square error reduction amount calculation unit (D calculation unit) 1205. Note that the video encoding apparatus of the fifth embodiment is basically configured as shown in FIG. 1, but an adaptive quantizer 1025 is provided instead of the adaptive quantizer 1021.
- the adaptive quantizer 1025 shown in FIG. 9 includes a coefficient level calculation unit 1201, an absolute value information bit number calculation unit (R abs calculation unit) 1206, a position / code number information bit number calculation unit (R locate_sign calculation unit) 1207, a high A cost coefficient level detection / removal unit 1204 and a square error reduction amount calculation unit (D calculation unit) 1205 are included.
- R abs calculation unit absolute value information bit number calculation unit
- R locate_sign calculation unit position / code number information bit number calculation unit
- D calculation unit square error reduction amount calculation unit
- the coefficient level calculation unit 1201 receives the orthogonal transform coefficient Kij and the quantization parameter QP, and outputs a temporary coefficient level L′ ij.
- the D calculation unit 1205 receives the quantization parameters QP and L′ ij and outputs D.
- D is a square error reduction amount with respect to Kij.
- the square error reduction amount D is calculated using the product of the square value of the quantization step Qs and the square value of L′ ij. .
- the R abs calculation unit 1206 receives L′ ij and outputs R abs which is the number of absolute value information bits.
- the R locate_sign calculation unit 1207 receives the orthogonal transform coefficient position (i, j) and outputs R locate_sign which is the number of position / code number information bits.
- the high cost coefficient level detection / removal unit 1204 inputs L′ ij, R abs , R locate_sign and D and outputs the coefficient level Lij.
- the high cost coefficient level detection / removal unit 1204 uses the slope ⁇ of D with respect to the sum of R locate_sign and R abs . In the case of uniform quantization, ⁇ is expressed by the above equation (4) using Qs.
- the high cost coefficient level detection / removal unit 1204 increases L'ij when the absolute value of L'ij is greater than 0 and the product of R locate and R value and ⁇ is greater than or equal to D.
- the cost coefficient level is determined and a value of 0 is output as Lij. If L'ij is not at a high cost coefficient level, L'ij is output as Lij.
- the adaptive quantizer 1025 performs transmission order ((3, 3) ⁇ (3, 2) ⁇ (2, for each orthogonal transform coefficient Kij (0 ⁇ i, j ⁇ 3) in each TU. 3) ⁇ (3, 1) ⁇ (2, 2) ⁇ (1, 3) ⁇ (3, 0) ⁇ (2, 1) ⁇ (1, 2) ⁇ (0, 3) ⁇ (2, 0) ⁇ (1, 1) ⁇ (0, 2) ⁇ (1, 0) ⁇ (0, 1) ⁇ (0, 0) in this order)
- step S401 the coefficient level calculation unit 1201 calculates a temporary coefficient level L′ ij corresponding to the orthogonal transform coefficient Kij using, for example, the above equation (1).
- step S402 If it is determined in step S402 that the value of L'ij is not 0, the process proceeds to step S403. If the value of L′ ij is 0, the process proceeds to step S407.
- step S403 the D calculation unit 1205 calculates D which is a square error reduction amount for the orthogonal transform coefficient.
- step S404 the R abs calculation unit 1206 calculates the number of bits R abs of the absolute value information of L′ ij.
- step S405 the R locate_sign calculation unit 1207 determines whether a non-zero Lij has already been output in the TU. If non-zero Lij is not yet output in the TU, the process proceeds to step S406. If non-zero Lij has already been output in the TU, R locate_sign is not calculated (the value of R locate_sign is set to 0), and the process proceeds to step S407.
- step S406 the R locate_sign calculation unit 1207 calculates the number of bits R locate_sign of the position / code number information determined based on the position of L′ ij.
- FIG. FIG. 11 is a block diagram illustrating a configuration of an adaptive quantizer 1026 in the video encoding device according to the sixth embodiment.
- Adaptive quantizer 1026 includes the maximum number of bits of coeff_sign_flag determined from the position of the significant coefficient level to be transmitted first in the number of position information bits.
- the adaptive quantizer 1026 uses a square error reduction amount with respect to Kij. Note that the video encoding apparatus of the sixth embodiment is basically configured as shown in FIG. 1, but an adaptive quantizer 1026 is provided instead of the adaptive quantizer 1021.
- Adaptive quantizer 11 1026 the coefficient level calculation unit 1201, the absolute value information bits calculation section (R abs calculating unit) 1206, a position / code number information bits calculation section (R locate_sign calculation unit) 1207, a high A cost coefficient level detection / removal unit 1204 and a square error reduction amount calculation unit (D calculation unit) 1205 are included.
- the coefficient level calculation unit 1201 receives the orthogonal transform coefficient Kij and the quantization parameter QP, calculates the temporary coefficient level L′ ij using, for example, the above equation (1), and outputs the temporary coefficient level L′ ij.
- the D calculation unit 1205 receives the quantization parameters QP, L′ ij, and Kij, and outputs D.
- D is a square error reduction amount with respect to Kij.
- the D calculation unit 1205 calculates D using Kij, Qs, and L′ ij as shown in the above equation (5).
- the R abs calculation unit 1206 receives L′ ij and outputs R abs which is the number of absolute value information bits.
- the R locate_sign calculation unit 1207 receives the orthogonal transform coefficient position (i, j) and outputs R locate_sign which is the number of bits of position / code number information.
- the high cost coefficient level detection / removal unit 1204 inputs L′ ij, R abs , R locate_sign , and D, and outputs the coefficient level Lij.
- the absolute value of L'ij is greater than 0, and the product of the sum and ⁇ of R Locate_sign and R abs is the case of the above D determines L'ij a high cost factor levels, Lij values of 0 Output as.
- the high cost coefficient level detection / removal unit 1204 outputs L'ij as Lij when L'ij is not at the high cost coefficient level.
- the D calculation unit 1205 calculates D using the quantization parameters QP and L′ ij, whereas in this embodiment, the D calculation unit 1205 includes Kij, Qs, and D is calculated using L′ ij, but the other processing is the same as in the fifth embodiment.
- the detection and removal of the high cost coefficient level is applied to all the non-zero temporary coefficient levels L′ ij, but a certain threshold value (quantization parameter QP or video property) Note that the temporary coefficient level L′ ij having an absolute value larger than 2) does not become a high cost coefficient level, for example, in order to reduce the amount of calculation.
- the detection and removal of the high cost coefficient level in each of the above embodiments may be applied only to the temporary coefficient level L′ ij.
- the temporary coefficient level L that has an absolute value equal to or less than the threshold value (for example, 2) Focusing on the fact that the number of bits of 'ij's coeff_abs_level_remaining ⁇ is expressed only by the TU code of the prefix part, in order to reduce the amount of calculation, the bit number of coeff_abs_level_remaining ⁇ of non-zero provisional coefficient level L'ij below the threshold is set The value of coeff_abs_level_greater1_flag and the value of coeff_abs_level_greater2_flag may be subtracted from the absolute value of L'ij.
- the threshold value for example, 2
- the adaptive quantizer applies the above-described detection and removal of the high cost coefficient level to all orthogonal transform coefficients.
- the high cost coefficient level is almost generated in the intra-predicted orthogonal transform coefficients.
- the adaptive quantizer may apply the above-described detection and removal of the high cost coefficient level only to the inter-predicted orthogonal transform coefficient.
- the orthogonal transform coefficient is a frequency-transformed prediction error image.
- transform_skip_flag shown in 7.3.9.11 Residual coding syntax of Non-Patent Document 1 is 1, that is, frequency conversion is performed in 4 ⁇ 4 TU.
- the adaptive quantizer of the present invention may be applied to a prediction error image to which transformation based on a unit matrix is applied.
- the position information is information indicating the positions of all significant coefficient levels of the TU (the sum of the number of bits of last_significant_x, last_significant_y, and siginificant_coeff_flag), but TU larger than 4 ⁇ 4 (8 ⁇ 8 TU , 16 ⁇ 16 TU, and 32 ⁇ 32 TU), in addition to the location information, 4 ⁇ 4 having 16 coefficient levels defined based on the description of 7.3.9.11 Residual coding syntax You may add the bit number of coded_sub_block_flag which shows the presence or absence of the significant coefficient level in a subblock.
- each of the above embodiments can be configured by hardware, it can also be realized by a computer program.
- the information processing system illustrated in FIG. 12 includes a processor 1001, a program memory 1002, a storage medium 1003 for storing video data, and a storage medium 1004 for storing a bitstream.
- the storage medium 1003 and the storage medium 1004 may be separate storage media, or may be storage areas composed of the same storage medium.
- a magnetic storage medium such as a hard disk can be used as the storage medium.
- the program memory 1002 stores a program for realizing the function of each block shown in FIG. Then, the processor 1001 implements the function of the video encoding device shown in FIG. 1B by executing processing according to the program stored in the program memory 1002.
- FIG. 13 is a block diagram showing the main part of the video encoding apparatus according to the present invention.
- the video encoding apparatus according to the present invention includes an orthogonal transform unit 11 that performs orthogonal transform on an image block to calculate orthogonal transform coefficients, and a quantization unit that quantizes the orthogonal transform coefficients to calculate a coefficient level. 12 and an entropy encoding unit 13 that entropy-encodes the position information of all the significant coefficient levels among the coefficient levels and then entropy-encodes the value information of each significant coefficient level and outputs a bitstream.
- the converting unit 12 includes a position information bit number calculating unit that calculates the number of position information bits determined based on the position of the first significant coefficient level in the transmission order included in the image block, and a significant coefficient level value information bit.
- a value information bit number calculation unit for calculating a number, and a high cost coefficient level in which a significant coefficient level satisfying a predetermined condition for the position information bit number and the value information bit number is 0
- a detection / removal unit A detection / removal unit.
- Orthogonal transformation means for orthogonally transforming an image block to calculate orthogonal transformation coefficients
- quantization means for quantizing the orthogonal transformation coefficients to calculate coefficient levels, and all significant coefficient levels among the coefficient levels
- Entropy-encoding means for entropy-encoding the value information of each significant coefficient level and outputting a bitstream after entropy-encoding the position information of each of the significant coefficient levels
- the quantization means includes: Position information bit number calculating means for calculating the number of position information bits determined based on the position of the first significant coefficient level in the transmission order included in the image block, and calculating the value information bit number of the significant coefficient level A value information bit number calculating means, and a significant coefficient level satisfying a predetermined condition between the position information bit number and the value information bit number is 0.
- That video coding apparatus which comprises a high cost factor level detection / removal means.
- the said quantization means contains the square error reduction amount calculation means which calculates the square error reduction amount with respect to an orthogonal transformation coefficient using a quantization parameter and the said significant coefficient level,
- the said high cost coefficient The level detection / removal means uses the number of position information bits, the number of value information bits, and the amount of square error reduction, and the square error reduction amount is obtained by calculating the position information bit number and the value information bit.
- the video encoding apparatus according to supplementary note 1 or supplementary note 2, wherein the significant coefficient level that is equal to or less than a value obtained by multiplying the sum by a predetermined multiplier is 0.
- the position information bit number calculation means indicates the position information bit number of the first significant coefficient level and the sign of all significant coefficient levels in the block as the position information bit number. The sum of the number of bits of information is calculated, and the value information bit number calculation means calculates the number of bits of information of the absolute value of the significant coefficient level as the value information bit number. Any video encoding device.
- the quantization means is any one of Supplementary notes 1 to 5 that uses the high-cost coefficient level detection / removal means only when the absolute value of the significant coefficient level is not more than a predetermined threshold value.
- Video encoding device is any one of Supplementary notes 1 to 5 that uses the high-cost coefficient level detection / removal means only when the absolute value of the significant coefficient level is not more than a predetermined threshold value.
- the value information bit number calculating means includes information indicating whether the significant coefficient level is greater than 1 from the absolute value information of the significant coefficient level greater than 2 included in the value information bit number.
- Supplementary note 10 The video of any one of Supplementary note 5 to Supplementary note 9, wherein the quantization means quantizes a prediction error image to which a transform based on a unit matrix is applied instead of a frequency transform to calculate a coefficient level. Encoding device.
- the position information bit number calculation means calculates the position information bit number in an image block (8 ⁇ 8 image block, 16 ⁇ 16 image block, or 32 ⁇ 32 image block) larger than 4 ⁇ 4.
- the video encoding device according to any one of appendix 5 to appendix 9, which includes the number of bits of information indicating presence / absence of a significant coefficient level in a 4 ⁇ 4 subblock having 16 coefficient levels.
- Orthogonal Transformer 12 Quantizer 13 Entropy Encoder 101 Frequency Converter 1020-1025 Adaptive Quantizer 103 Entropy Encoder 104 Inverse Frequency Transform / Inverse Quantizer 105 Buffer 106 Predictor 107 Estimator 1001 Processor 1002 Program Memory 1003 Storage medium 1004 Storage medium 1201 Coefficient level calculation part 1202 Value information bit number calculation part (R value calculation part) 1203 Location information bit number calculation part (R locate calculation part) 1204 High cost coefficient level detection / removal unit 1205 Square error reduction amount calculation unit (D calculation unit) 1206 Absolute value information bit number calculation part (R abs calculation part) 1207 Position / code number information bit number calculation part (R locate_sign calculation part)
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Algebra (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
図1 は、本発明による映像符号化装置の第1の実施形態を示すブロック図である。図1 を参照して、ディジタル化された映像の各フレームを入力画像としてビットストリームを出力する第1 の実施形態の映像符号化装置の構成を説明する。
上記した第1の実施形態では、高コスト係数レベルを正確に検出できない場合がある。そこで、高コスト係数レベルをより正確に検出するために、第2の実施形態では、図4 に示す構成の適応量子化器1022を用いる。なお、第2の実施形態の映像符号化装置は、基本的に図1に示されるように構成されるが、適応量子化器1021に代えて、適応量子化器1022が設けられる。
ただし、αは、-f <α<1-f を満たす係数である。
第2の実施形態では、Kij を用いずに2乗誤差削減量を計算するために、正確に高コスト係数レベルを検出できない場合がある。そこで、高コスト係数レベルをより正確に検出するために、第3の実施形態では、図6 に示す構成の適応量子化器1023を用いる。なお、第3の実施形態の映像符号化装置は、基本的に図1に示されるように構成されるが、適応量子化器1021に代えて、適応量子化器1023が設けられる。
図7 は、第4の実施形態の映像符号化装置における適応量子化器1024の構成を示すブロック図である。適応量子化器1024は、最初に伝送する有意係数レベルの位置から決定されるcoeff_sign_flag の最大ビット数を位置情報ビット数に含める。なお、第4の実施形態の映像符号化装置は、基本的に図1に示されるように構成されるが、適応量子化器1021に代えて、適応量子化器1024が設けられる。
図9 は、第5の実施形態の映像符号化装置における適応量子化器1025の構成を示すブロック図である。適応量子化器1025は、最初に伝送する有意係数レベルの位置から決定されるcoeff_sign_flag の最大ビット数を位置情報ビット数に含める。また、適応量子化器1025は、2乗誤差削減量計算部(D 計算部)1205によって算出されたKij に対する2乗誤差削減量を用いる。なお、第5の実施形態の映像符号化装置は、基本的に図1に示されるように構成されるが、適応量子化器1021に代えて、適応量子化器1025が設けられる。
図11は、第6の実施形態の映像符号化装置における適応量子化器1026の構成を示すブロック図である。適応量子化器1026は、最初に伝送する有意係数レベルの位置から決定されるcoeff_sign_flag の最大ビット数を位置情報ビット数に含める。また、適応量子化器1026は、Kij に対する2乗誤差削減量を用いる。なお、第6の実施形態の映像符号化装置は、基本的に図1に示されるように構成されるが、適応量子化器1021に代えて、適応量子化器1026が設けられる。
12 量子化部
13 エントロピー符号化部
101 周波数変換器
1020~1025 適応量子化器
103 エントロピー符号化器
104 逆周波数変換/逆量子化器
105 バッファ
106 予測器
107 推定器
1001 プロセッサ
1002 プログラムメモリ
1003 記憶媒体
1004 記憶媒体
1201 係数レベル計算部
1202 値情報ビット数計算部(Rvalue 計算部)
1203 位置情報ビット数計算部(Rlocate 計算部)
1204 高コスト係数レベル検出/除去部
1205 2乗誤差削減量計算部(D 計算部)
1206 絶対値情報ビット数計算部(Rabs計算部)
1207 位置/符号個数情報ビット数計算部(Rlocate_sign計算部)
Claims (10)
- 画像ブロックを直交変換して直交変換係数を計算する直交変換手段と、
前記直交変換係数を量子化して係数レベルを計算する量子化手段と、
前記係数レベルのうちすべての有意係数レベルの位置情報をエントロピー符号化した後に、それぞれの前記有意係数レベルの値情報をエントロピー符号化してビットストリームを出力するエントロピー符号化手段とを備える映像符号化装置であって、
前記量子化手段は、
前記画像ブロックに含まれる伝送順で最初に有意な係数レベルの位置に基づいて決定される位置情報ビット数を計算する位置情報ビット数計算手段と、
有意な係数レベルの値情報ビット数を計算する値情報ビット数計算手段と、
前記位置情報ビット数と前記値情報ビット数とが所定の条件を満たす有意な係数レベルを0とする高コスト係数レベル検出/除去手段とを含む
ことを特徴とする映像符号化装置。 - 前記高コスト係数レベル検出/除去手段は、前記位置情報ビット数が前記値情報ビット数よりも大きい場合に、有意な係数レベルを0とする
請求項1記載の映像符号化装置。 - 前記量子化手段は、量子化パラメータと前記有意な係数レベルとを用いて、直交変換係数に対する2乗誤差削減量を計算する2乗誤差削減量計算手段を含み、
前記高コスト係数レベル検出/除去手段は、前記位置情報ビット数、前記値情報ビット数、および前記2乗誤差削減量を用いて、前記2乗誤差削減量が、前記位置情報ビット数と前記値情報ビットとの和に対して所定の乗数を乗じた値以下となる前記有意な係数レベルを0とする
請求項1または請求項2記載の映像符号化装置。 - 前記2乗誤差削減量計算手段は、前記量子化パラメータ、前記有意係数レベル、および前記直交変換係数を用いて、前記2乗誤差削減量を計算する
請求項3記載の映像符号化装置。 - 画像ブロックを直交変換して直交変換係数を計算し、
前記直交変換係数を量子化して係数レベルを計算し、
前記係数レベルのうちすべての有意係数レベルの位置情報をエントロピー符号化した後に、それぞれの前記有意係数レベルの値情報をエントロピー符号化してビットストリームを出力する映像符号化方法であって、
前記係数レベルを計算する際に、
前記画像ブロックに含まれる伝送順で最初に有意な係数レベルの位置に基づいて決定される位置情報ビット数を計算し、
有意な係数レベルの値情報ビット数を計算し、
前記位置情報ビット数と前記値情報ビット数とが所定の条件を満たす有意な係数レベルを0とする
ことを特徴とする映像符号化方法。 - 前記所定の条件を、前記位置情報ビット数が前記値情報ビット数よりも大きいことであるとする
請求項5記載の映像符号化方法。 - 前記係数レベルを計算する際に、量子化パラメータと前記有意な係数レベルとを用いて、直交変換係数に対する2乗誤差削減量を計算し、
前記位置情報ビット数、前記値情報ビット数、および前記2乗誤差削減量を用いて、前記2乗誤差削減量が、前記位置情報ビット数と前記値情報ビットとの和に対して所定の乗数を乗じた値以下となる前記有意な係数レベルを0とする
請求項5または請求項6記載の映像符号化方法。 - 前記量子化パラメータ、前記有意係数レベル、および前記直交変換係数を用いて、前記2乗誤差削減量を計算する
請求項7記載の映像符号化方法。 - コンピュータに、
画像ブロックを直交変換して直交変換係数を計算する処理と、
前記直交変換係数を量子化して係数レベルを計算する処理と、
前記係数レベルのうちすべての有意係数レベルの位置情報をエントロピー符号化した後に、それぞれの前記有意係数レベルの値情報をエントロピー符号化してビットストリームを出力する処理とを実行させ、
前記係数レベルを計算する処理で、
前記画像ブロックに含まれる伝送順で最初に有意な係数レベルの位置に基づいて決定される位置情報ビット数を計算する処理と、
有意な係数レベルの値情報ビット数を計算する処理と、
前記位置情報ビット数と前記値情報ビット数とが所定の条件を満たす有意な係数レベルを0とする処理とを
実行させるための映像符号化プログラム。 - コンピュータに、
前記位置情報ビット数が前記値情報ビット数よりも大きい場合に、有意な係数レベルを0とする処理を実行させるための請求項9記載の映像符号化プログラム。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2014515952A JP5574072B1 (ja) | 2012-12-27 | 2013-12-03 | 映像符号化装置、映像符号化方法および映像符号化プログラム |
BR112015015122A BR112015015122A2 (pt) | 2012-12-27 | 2013-12-03 | dispositivo de codificação de vídeo, método de codificação de vídeo e programa de codificação de vídeo |
US14/651,466 US10003804B2 (en) | 2012-12-27 | 2013-12-03 | Video coding device using quantizing an orthogonal transform coefficient |
EP13868141.6A EP2941000B1 (en) | 2012-12-27 | 2013-12-03 | Video coding device, video coding method, and video coding program |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2012285238 | 2012-12-27 | ||
JP2012-285238 | 2012-12-27 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2014103182A1 true WO2014103182A1 (ja) | 2014-07-03 |
Family
ID=51020309
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2013/007080 WO2014103182A1 (ja) | 2012-12-27 | 2013-12-03 | 映像符号化装置、映像符号化方法および映像符号化プログラム |
Country Status (5)
Country | Link |
---|---|
US (1) | US10003804B2 (ja) |
EP (1) | EP2941000B1 (ja) |
JP (1) | JP5574072B1 (ja) |
BR (1) | BR112015015122A2 (ja) |
WO (1) | WO2014103182A1 (ja) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015045736A1 (ja) * | 2013-09-30 | 2015-04-02 | 日本放送協会 | 画像符号化装置、画像復号装置及びそれらのプログラム |
US9930348B2 (en) * | 2014-03-14 | 2018-03-27 | Qualcomm Incorporated | Coefficient level coding in a video coding process |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2010087771A (ja) * | 2008-09-30 | 2010-04-15 | Nec Electronics Corp | 画像符号化装置及び方法 |
JP2011509642A (ja) * | 2008-01-08 | 2011-03-24 | クゥアルコム・インコーポレイテッド | Cabacコーダのためのレート歪みモデリングに基づいた量子化 |
JP2011176483A (ja) * | 2010-02-23 | 2011-09-08 | Oki Electric Industry Co Ltd | 量子化装置、プログラム及び方法、並びに、動画像符号化装置 |
WO2012075193A1 (en) * | 2010-12-03 | 2012-06-07 | Qualcomm Incorporated | Coding the position of a last significant coefficient within a video block based on a scanning order for the block in video coding |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6819803B2 (en) * | 2001-07-02 | 2004-11-16 | International Business Machines Corporation | Faster lossless rotation of JPEG images |
US20120207400A1 (en) * | 2011-02-10 | 2012-08-16 | Hisao Sasai | Image coding method, image coding apparatus, image decoding method, image decoding apparatus, and image coding and decoding apparatus |
KR101598142B1 (ko) * | 2011-10-31 | 2016-02-26 | 삼성전자주식회사 | 변환 계수 레벨의 엔트로피 부호화 및 복호화를 위한 컨텍스트 모델 결정 방법 및 장치 |
-
2013
- 2013-12-03 EP EP13868141.6A patent/EP2941000B1/en active Active
- 2013-12-03 JP JP2014515952A patent/JP5574072B1/ja not_active Expired - Fee Related
- 2013-12-03 US US14/651,466 patent/US10003804B2/en not_active Expired - Fee Related
- 2013-12-03 BR BR112015015122A patent/BR112015015122A2/pt not_active IP Right Cessation
- 2013-12-03 WO PCT/JP2013/007080 patent/WO2014103182A1/ja active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2011509642A (ja) * | 2008-01-08 | 2011-03-24 | クゥアルコム・インコーポレイテッド | Cabacコーダのためのレート歪みモデリングに基づいた量子化 |
JP2010087771A (ja) * | 2008-09-30 | 2010-04-15 | Nec Electronics Corp | 画像符号化装置及び方法 |
JP2011176483A (ja) * | 2010-02-23 | 2011-09-08 | Oki Electric Industry Co Ltd | 量子化装置、プログラム及び方法、並びに、動画像符号化装置 |
WO2012075193A1 (en) * | 2010-12-03 | 2012-06-07 | Qualcomm Incorporated | Coding the position of a last significant coefficient within a video block based on a scanning order for the block in video coding |
Non-Patent Citations (1)
Title |
---|
"High efficiency video coding (HEVC) text specification draft 9", JCTVC-K1003 V9, JOINT COLLABORATIVE TEAM ON VIDEO CODING (JCT-VC) OF ITU-T SG16 WP3 AND ISO/IEC JTC1/SC29/WG11 LLTH MEETING, October 2012 (2012-10-01) |
Also Published As
Publication number | Publication date |
---|---|
US10003804B2 (en) | 2018-06-19 |
JPWO2014103182A1 (ja) | 2017-01-12 |
EP2941000A1 (en) | 2015-11-04 |
US20150319440A1 (en) | 2015-11-05 |
JP5574072B1 (ja) | 2014-08-20 |
EP2941000B1 (en) | 2020-11-04 |
EP2941000A4 (en) | 2016-06-08 |
BR112015015122A2 (pt) | 2017-07-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11943485B2 (en) | Method of coding and decoding images, coding and decoding device and computer programs corresponding thereto | |
JP5537681B2 (ja) | 変換ユニット内の複数サインビット秘匿 | |
US11647205B2 (en) | Video encoding device, video decoding device, video encoding method, video decoding method, and program using inter prediction | |
CN107396116B (zh) | 视频编码和解码设备及非暂时性计算机可读存储介质 | |
US9609342B2 (en) | Compression for frames of a video signal using selected candidate blocks | |
WO2018123313A1 (ja) | 映像符号化方法、映像復号方法、映像符号化装置、映像復号装置及びプログラム | |
US9838681B2 (en) | Apparatus of decoding video data | |
KR20140064972A (ko) | 화상 부호화 및 복호 방법, 장치, 프로그램 | |
WO2015008417A1 (ja) | 映像符号化装置、映像復号装置、映像符号化方法、映像復号方法及びプログラム | |
JP5574072B1 (ja) | 映像符号化装置、映像符号化方法および映像符号化プログラム | |
JP5644987B1 (ja) | 映像符号化装置、映像符号化方法および映像符号化プログラム | |
JP4451759B2 (ja) | 可逆ビデオ符号化装置,可逆ビデオ符号化方法,可逆ビデオ復号装置,可逆ビデオ復号方法,可逆ビデオ符号化プログラム,可逆ビデオ復号プログラムおよびそれらのプログラムの記録媒体 | |
KR20180040827A (ko) | 부호화 유닛의 그룹을 사용하는 비디오 코딩 방법 및 장치 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2014515952 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 13868141 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2013868141 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14651466 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
REG | Reference to national code |
Ref country code: BR Ref legal event code: B01A Ref document number: 112015015122 Country of ref document: BR |
|
ENP | Entry into the national phase |
Ref document number: 112015015122 Country of ref document: BR Kind code of ref document: A2 Effective date: 20150623 |