WO2011048903A1 - 動画像符号化装置、動画像復号装置、および、データ構造 - Google Patents
動画像符号化装置、動画像復号装置、および、データ構造 Download PDFInfo
- Publication number
- WO2011048903A1 WO2011048903A1 PCT/JP2010/066245 JP2010066245W WO2011048903A1 WO 2011048903 A1 WO2011048903 A1 WO 2011048903A1 JP 2010066245 W JP2010066245 W JP 2010066245W WO 2011048903 A1 WO2011048903 A1 WO 2011048903A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- partition
- prediction
- vector candidate
- prediction vector
- candidate
- Prior art date
Links
- 239000013598 vector Substances 0.000 claims abstract description 1074
- 238000005192 partition Methods 0.000 claims abstract description 569
- 238000004364 calculation method Methods 0.000 claims description 38
- 230000002123 temporal effect Effects 0.000 description 96
- 230000009467 reduction Effects 0.000 description 51
- 230000004048 modification Effects 0.000 description 50
- 238000012986 modification Methods 0.000 description 50
- 238000000034 method Methods 0.000 description 25
- 238000010586 diagram Methods 0.000 description 14
- 238000013139 quantization Methods 0.000 description 13
- 230000000694 effects Effects 0.000 description 10
- 230000007423 decrease Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000009466 transformation Effects 0.000 description 3
- 229910020896 PMVx Inorganic materials 0.000 description 2
- 229910020894 PMVy Inorganic materials 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000006837 decompression Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000002730 additional effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
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
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/553—Motion estimation dealing with occlusions
-
- 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
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/56—Motion estimation with initialisation of the vector search, e.g. estimating a good candidate to initiate a search
-
- 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
- H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
Definitions
- the present invention relates to a moving picture coding apparatus that codes a moving picture and generates coded data.
- the present invention also relates to a moving picture decoding apparatus that decodes encoded data generated using such a moving picture encoding apparatus.
- a moving picture coding apparatus is used to efficiently transmit or record moving pictures. Also, motion-compensated prediction using a motion vector is used for coding a moving image in a moving-image coding apparatus.
- H264 / AVC etc. are mentioned as a moving image coding system using motion compensation prediction.
- an input moving image is divided into a plurality of partitions for each frame, and the left side of a partition to be encoded (hereinafter referred to as a "target partition") among the plurality of partitions.
- a target partition the left side of a partition to be encoded
- Estimate the prediction vector to be assigned to the target partition using the median (median) of the motion vector assigned to each of the adjacent partition, the partition adjacent to the upper side of the target partition, and the upper right partition of the target partition.
- Non-Patent Document 1 as a technique for encoding a motion vector more efficiently, in a frame before a frame including a partition to be encoded, co-locate is a partition that occupies the same position as the target partition.
- the candidate of the prediction vector (referred to as “temporal prediction vector candidate”) to be allocated to the target partition using the motion vector allocated to the partition and the median of the motion vector allocated to each of the partitions around the co-located partition
- a prediction vector with good coding efficiency among the candidate and a prediction vector candidate (referred to as “spatial prediction vector candidate”) estimated based on the technology described as the above-mentioned conventional general technology.
- MV Com with candidate as prediction vector Technique called etition is disclosed.
- Non-Patent Document 1 even when the accuracy of the temporal motion vector is not high, that is, even when the temporal prediction vector candidate becomes dissimilar to the motion vector allocated to the target partition, prediction is performed. It is used as one of vector candidates.
- MV Competition a flag indicating which prediction vector candidate is selected as a prediction vector to be used for encoding from a plurality of prediction vector candidates is encoded, and therefore, when using a motion vector candidate with low accuracy, the code amount of the flag As a result, the coding efficiency may be reduced.
- the prediction accuracy of the above temporally predicted vector candidate tends to decrease. Such a tendency is not used.
- the present invention has been made in view of the above problems, and an object thereof is to increase the accuracy of prediction vector candidates by narrowing down prediction vector candidates without requiring a flag among a plurality of prediction vector candidates. This is to realize a moving picture coding apparatus with high coding efficiency.
- the moving picture coding apparatus is configured to calculate a difference between a motion vector and a prediction vector allocated to each of a plurality of partitions obtained by dividing a frame constituting a moving picture.
- the predicted vector is adjacent to the left side of the target partition, the upper partition adjacent to the upper side of the target partition, and the right side of the upper partition.
- First target vector candidate which is the median of the motion vector allocated to each of the upper right partitions, or the second target vector candidate which is the motion vector allocated to the left partition, or the above target in the decoded frame Coloke occupying the same position as partition
- third prediction vector candidates which are motion vectors allocated to the target partition, among any one of the prediction vector candidates selected according to the prediction accuracy of the third prediction vector candidate, according to the coding cost It is characterized in that it is any one prediction vector selected.
- the moving picture coding apparatus configured as described above includes the second prediction according to the prediction accuracy of the third prediction vector candidate corresponding to the temporal prediction vector candidate in Non-Patent Document 1. Any one of the vector candidate and the third prediction vector candidate is selected, and the first corresponding to the spatial prediction vector candidate in Non-Patent Document 1 is selected according to the coding cost.
- One of the prediction vector candidates or the prediction vector selected from the second prediction vector candidate or the third prediction vector candidate is selected and set as the prediction vector to be allocated to the target partition Accuracy of the third predicted vector candidate corresponding to the temporally predicted vector candidate in Non-Patent Document 1 Even in the case of a decrease, even if the coding cost of the first prediction vector candidate corresponding to the spatial prediction vector candidate in Non-Patent Document 1 increases, coded data with high coding efficiency is obtained. The effect is that it can be generated.
- the moving picture decoding apparatus encodes, together with the moving picture, a difference vector between a motion vector and a prediction vector allocated to each of a plurality of partitions obtained by dividing a frame constituting the moving picture.
- a moving image decoding apparatus for decoding encoded data obtained in the above process, the left partition adjacent to the left side of the target partition, the upper partition adjacent to the upper side of the target partition, and the upper right partition adjacent to the right side of the upper partition First calculating means for setting the median of the motion vector assigned to each of the first prediction vector candidates as candidates for the prediction vector to be assigned to the target partition, and the motion vector assigned to the left partition Prediction vector candidate to be assigned to partition
- Calculation means for setting a second predicted vector candidate, and a candidate for a predicted vector for allocating to the target partition a motion vector allocated to a co-located partition occupying the same position as the target partition in a decoded frame Among the second predicted vector candidate or the third predicted vector candidate according to the third calculation means for setting the third predicted vector candidate
- the second predicted vector candidate or the third predicted vector candidate is selected according to the prediction accuracy of the third predicted vector candidate. Since the first selection means for selecting any one of the prediction vector candidates is provided, the second prediction vector candidate or the second prediction vector candidate may be selected according to the prediction accuracy of the third prediction vector candidate. Among the three predicted vector candidates, any one predicted vector candidate can be selected.
- the moving picture decoding apparatus configured as described above refers to the flag included in the encoded data and selects the first predicted vector candidate or the first selection means. And second selecting means for selecting any one of the predicted vector candidates and setting the selected candidate vector as the predicted vector to be allocated to the target partition, so that the prediction of the third predicted vector candidate is performed. Even in the case where the accuracy decreases and the coding cost in the case of using the first prediction vector candidate increases, the coded data generated with high coding efficiency can be decoded. It plays an effect.
- the prediction vector has a left partition adjacent to the left side of the target partition, an upper partition adjacent to the upper side of the target partition, and an upper right adjacent to the right side of the upper partition.
- the same position as the target partition in the first predicted vector candidate which is the median of the motion vector allocated to each of the partitions, or the second predicted vector candidate or decoded frame which is the motion vector allocated to the left partition Occupying Among the third prediction vector candidates that are motion vectors allocated to the partition, among any one of the prediction vector candidates selected according to the prediction accuracy of the third prediction vector candidate, according to the coding cost It is characterized in that it is any one selected prediction vector.
- any one of the second predicted vector candidate and the third predicted vector candidate is selected as the predicted vector candidate.
- And can be selected according to the prediction accuracy of the third prediction vector candidate. Therefore, according to the above configuration, it is possible to realize encoded data with high decoding efficiency.
- a difference vector between a motion vector and a prediction vector allocated to each of a plurality of partitions obtained by dividing a frame constituting the moving picture is the moving picture
- the prediction vector includes a left partition adjacent to the left side of the target partition, an upper partition adjacent to the upper side of the target partition, and an upper right partition adjacent to the right side of the upper partition.
- FIG. 1 is a block diagram showing a configuration of a video encoding device according to an embodiment. It is a figure for demonstrating the operation
- A) shows a motion vector allocated to each of the partitions around the target partition and a prediction vector allocated to the target partition.
- (B) shows a target partition, a prediction vector assigned to the target partition, a co-located partition, and a motion vector assigned to the co-located partition.
- FIG. 17 is a view for explaining an operation of a motion vector redundancy reduction unit in a moving picture decoding apparatus according to a third modified example of the embodiment, and showing a GOP structure configured by hierarchy B pictures.
- the configuration of the moving picture coding apparatus 1 according to the present embodiment will be described with reference to FIGS. 1 to 5.
- FIG. 2 is a block diagram showing the configuration of the moving picture coding apparatus 1.
- the moving picture coding apparatus 1 includes a transform / quantization unit 11, a variable length coding unit 12, an inverse quantization / inverse transform unit 13, a buffer memory 14, an intra predicted image generation unit 15, and prediction.
- An image generation unit 16 a motion vector estimation unit 17, a prediction scheme control unit 18, a motion vector redundancy reduction unit 19, an adder 21, and a subtractor 22 are provided.
- An input image # 1 is sequentially input to the video encoding device 1.
- the input image # 1 is an image signal corresponding to each frame of video data, and is an image signal corresponding to each frame of a 60 Hz progressive signal, for example.
- the moving picture coding apparatus 1 performs coding processing of an input image # 1 and outputs coded data # 2.
- the transformation / quantization unit 11 uses the input image # 1 divided into block images (hereinafter referred to as “macroblocks”) composed of a plurality of adjacent pixels, and the prediction output from the prediction method control unit 18 described later.
- the differential image # 22 with the image # 18a is converted to a frequency component by DCT (Discrete Cosine Transform) conversion, and then the frequency component is quantized to generate quantized prediction residual data # 11.
- the quantization is an operation of associating the frequency component with an integer value.
- the macro block to be processed is referred to as a “target macro block”.
- the size of the macro block is, for example, 16 ⁇ 16 pixels, but the present invention is not limited by the specific size of the macro block.
- the inverse quantization / inverse transform unit 13 decodes the quantized prediction residual data # 11 to generate a prediction residual # 13. Specifically, the inverse quantization / inverse transform unit 13 inversely quantizes the quantized prediction residual data # 11, that is, associates the integer values constituting the quantized prediction residual data # 11 with frequency components. And inverse DCT transform of the frequency component, that is, the target macroblock based on the frequency component is decoded to generate a prediction residual # 13.
- the adder 21 adds the prediction residual # 13 and the predicted image # 18a to generate a decoded image # 21.
- the generated decoded image # 21 is stored in the buffer memory 14.
- the intra prediction image generation unit 15 extracts the local decoded image # 14a from the decoded image # 21 stored in the buffer memory 14, performs intraframe prediction based on the local decoded image # 14a, and generates an intra prediction image # 15. Do.
- the size of the local decoded image # 14a is, for example, 16 ⁇ 16 pixels, 8 ⁇ 8 pixels, or 4 ⁇ 4 pixels, but the present invention is limited by the specific size of the local decoded image # 14a is not.
- the motion vector estimation unit 17 divides the target macroblock into one or more partitions, and sequentially assigns a motion vector to each partition. Specifically, based on the input image # 1 and an image (hereinafter referred to as a reference image # 14b) in which the entire frame has already been decoded and stored in the buffer memory 14, the motion vector estimation unit 17 Among the partitions, a motion vector # 17 is calculated and allocated to a partition to be processed (hereinafter referred to as "target partition"). Further, relative position information (hereinafter referred to as reference image relative position) of a reference image with reference to a frame to which each partition belongs is calculated and allocated. The calculated motion vector # 17 is output to the predicted image generation unit 16 and the motion vector redundancy reduction unit 19, and is stored in the buffer memory 14.
- the size of the partition is 16 ⁇ 16 pixels, 16 ⁇ 8 pixels, 8 ⁇ 16 pixels, 8 ⁇ 8 pixels, 8 ⁇ 4 pixels, 4 ⁇ 8 pixels, or 4 ⁇ 4 pixels.
- the invention is not limited by the specific partition size.
- the motion vector estimation unit 17 selects, from among the sizes of the partitions forming the target macroblock, the size of the partition with the lowest coding cost.
- RD cost and M cost are mentioned as an example of the above-mentioned coding cost.
- D represents the residual between the locally decoded image and the input image
- R represents the code amount of the inter prediction parameter .
- the coefficient ⁇ may be a constant or may be a function of a quantization parameter that controls the roughness of the quantization.
- the motion vector estimation unit 17 outputs the size of the target partition and an index indicating the image number of the frame in which the target partition is included. The motion vector and the index are determined for each target partition and output.
- the motion vector estimation unit 17 determines whether to apply a coding mode called skip mode to the target partition.
- the skip mode is a mode in which the quantized prediction residual data # 11 is not coded.
- the motion vector estimation unit 17 has a coding cost that is higher than the skip threshold, which is a predetermined threshold. When it is smaller, the skip mode is applied to the target partition (target macro block).
- the prediction image generation unit 16 performs motion compensation based on the motion vector # 17 for each partition on the reference image # 14b stored in the buffer memory 14 to generate an inter prediction image # 16.
- the prediction method control unit 18 compares the intra prediction image # 15, the inter prediction image # 16, and the input image # 1, and selects one of the intra prediction image # 15 and the inter prediction image # 16. Is selected and output as a predicted image # 18a. Further, the prediction method control unit 18 outputs a prediction mode # 18b, which is information indicating which one of the intra prediction image # 15 and the inter prediction image # 16 is selected. The predicted image # 18a is input to the subtractor 22.
- the prediction mode # 18 b is stored in the buffer memory 14 and input to the variable-length coding unit 12.
- the motion vector redundancy reduction unit 19 is allocated to another partition before the motion vector # 17 is allocated to the target partition in the motion vector estimation unit 17, and based on the motion vector group # 14 c stored in the buffer memory 14. Calculate the prediction vector. Also, the motion vector redundancy reduction unit 19 takes the difference between the prediction vector and the motion vector # 17 to generate a difference motion vector # 19a. The generated differential motion vector # 19a is output to the variable-length coding unit 12. In addition, when there are a plurality of prediction vectors, the motion vector redundancy reduction unit 19 outputs a flag # 19 b indicating which prediction vector was used to generate the differential motion vector # 19 a. The details of the motion vector redundancy reduction unit 19 will not be described here because they will be described in detail below.
- variable-length coding unit 12 performs variable-length coding on the quantized prediction residual data # 11, the difference motion vector # 19a, the prediction mode # 18b, and the flag # 19b, and generates the coded data # 2 Do. However, when the skip mode is applied, the variable-length coding unit 12 does not encode the quantized prediction residual data # 11.
- variable-length coding unit 12 embeds information indicating whether or not the skip mode is applied in block mode information which is a header of the coded data # 2.
- the subtractor 22 obtains the difference between the input image # 1 and the predicted image # 18a, and outputs a difference image # 22.
- the prediction vector to be calculated with reference to a spatial motion vector group consisting of motion vectors allocated to encoded partitions around the target partition in the target frame is hereinafter referred to as a spatial prediction vector.
- the left partition adjacent to the left side of the target partition, the upper partition adjacent to the upper side of the target partition, and the right side of the upper partition as one of methods for calculating the spatial prediction vector with reference to the above spatial motion vector group
- the motion vector assigned to the left partition adjacent to the left side of the target partition may be used,
- the motion vector assigned to the upper partition adjacent to the upper edge of the target partition may be used,
- An average of motion vectors allocated to each of the left partition adjacent to the left side of the target partition, the upper partition adjacent to the upper side of the target partition, and the upper right partition adjacent to the right side of the upper partition may be used.
- a motion vector with the largest difference between the motion vector and the median may be used.
- Temporal prediction vector Calculated with reference to a temporal motion vector group including a partition occupying the same position as the target partition in the encoded frame (hereinafter referred to as a co-located partition) and a motion vector allocated to a partition around the co-located partition
- the prediction vector is hereinafter referred to as a temporal prediction vector.
- One of the methods of calculating a temporal prediction vector with reference to the above-mentioned temporal motion vector group is a method of using a motion vector of a co-located partition which occupies the same position as a target partition in a coded frame.
- the median of motion vectors assigned to each of the left partition adjacent to the left side of the co-located partition, the upper partition adjacent to the upper side of the co-located partition, and the upper right partition adjacent to the right side of the upper partition may be used. You may use the average of From the co-located partition, it is possible to use a motion vector allocated to a spatially displaced position partition by the median of the spatial prediction vector group or the average motion vector.
- the method of calculating the temporal prediction vector is not limited to the method described above.
- Motion vector redundancy reduction unit 19 Motion vector redundancy reduction unit 19
- the configuration of the motion vector redundancy reduction unit will be described with reference to FIGS. 1 and 3A to 3B.
- FIG. 1 is a block diagram showing the configuration of the motion vector redundancy reduction unit 19.
- the motion vector redundancy reduction unit 19 includes a prediction vector candidate generation unit 31, a prediction vector selection unit 32, a PMV flag generation unit 33, and a subtractor 34.
- the predicted motion vector candidate generation unit 31 calculates motion vector candidates, which are motion vector candidates to be assigned to the target partition, based on the motion vector group # 14c stored in the buffer memory. Further, the prediction vector candidate generation unit 31 outputs a first spatial prediction vector candidate # 311 and a second prediction vector candidate # 314. The configuration of the prediction vector candidate generation unit 31 will not be described here because it will be described in detail below.
- the second prediction vector candidate # 314 is a prediction vector selected from the spatial prediction vector candidate and the temporal prediction vector candidate in consideration of the prediction accuracy of the temporal prediction vector candidate, and, semantically, It can also be called a space-time selection prediction vector.
- the selection of the second predicted vector candidate # 314 is selected by the same method in the encoding device and the decoding device, and is not switched by the flag 19b added to the partition.
- the number of secondary prediction vector candidates # 314 is described as 1, the number of secondary prediction vector candidates # 314 may be more than one as shown in the supplementary item 1 described later.
- the prediction vector selection unit 32 compares the first spatial prediction vector candidate # 311 output from the prediction vector candidate generation unit 31 with the second prediction vector candidate # 314, and selects any one prediction vector candidate. It selects and sets to prediction vector # 32a allocated to an object partition.
- the prediction vector selection unit 32 calculates the coding cost for each of the first spatial prediction vector candidate # 311 and the second prediction vector candidate # 314, and the coding A prediction vector candidate with a smaller cost is selected and set as a prediction vector # 32a.
- the prediction vector selection unit 32 is selection information # which is information indicating which prediction vector candidate has been selected among the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314. 32 b is generated and output to the PMV flag generation unit 33.
- the PMV flag generation unit 33 selects any prediction vector candidate by the prediction vector selection unit 32 among the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314 based on the selection information # 32b. Is generated to indicate whether or not is selected.
- the generated flag # 19 b is output to the variable length coding unit 12.
- the prediction vector selection unit 32 selects the prediction vector # 32a from the two prediction vector candidates (the first spatial prediction vector candidate # 311 and the second prediction vector candidate # 314). Therefore, the flag # 19 b is a 1-bit flag.
- the flag # 19 b is a flag of 2 bits or more. is there.
- the flag # 19b may be coded by fixed-length coding or may be coded using arithmetic coding.
- the PMV flag generation unit 33 is shown as a means different from the variable length coding unit 12 in order to explicitly show the encoding of the flag # 19 b.
- the selection information # 32b may be directly encoded.
- the PMV flag generator 33 is positioned to be included in the variable-length coding unit 12.
- a subtractor 34 is a difference motion vector that is a difference vector between the prediction vector # 32a assigned to the target partition set by the prediction vector selection unit 32 and the motion vector # 17 assigned to the target partition by the motion vector estimation unit 17 Generate 19a.
- the generated differential motion vector is output to the variable length coding unit 12.
- the prediction vector candidate generation unit 31 includes a first spatial prediction vector calculation unit 311, a second spatial prediction vector calculation unit 312, a temporal prediction vector calculation unit 313, and a prediction vector candidate selection unit. It has 314.
- the first spatial prediction vector calculation unit 311 is adjacent to the upper side of the target block, the motion vector allocated to the partition adjacent to the left side of the target block, the motion vector allocated to the partition adjacent to the upper side of the target block, and the upper side
- the first spatial prediction vector candidate # 311 is calculated by taking the median with the motion vector allocated to the partition adjacent to the right side of the partition (hereinafter, referred to as “upper right partition of target partition”).
- the first spatial prediction vector calculation unit 311 calculates the motion vector MVa allocated to the partition A adjacent to the left side of the target block and the partition adjacent to the upper side of the target block.
- the first spatial prediction vector candidate # 311 is calculated.
- the median is an arithmetic operation which takes the median value of elements
- the median of a vector is a vector which takes the median value for each corresponding component.
- the motion vector assigned to partition A is (MVax, MVay)
- the motion vector assigned to partition B is (MVbx, MVby)
- the motion vector assigned to partition C is (MVcx, MVcy)
- median means taking the median value of the elements in parentheses.
- the second spatial prediction vector calculation unit 312 sets the motion vector allocated to the partition adjacent to the left side of the target partition as the second spatial prediction vector candidate # 312. That is, the second spatial prediction vector calculation unit 312 sets the motion vector MVa allocated to the partition A in (a) of FIG. 3 as the second spatial prediction vector candidate # 312.
- the temporal prediction vector calculation unit 313 assigns a motion vector allocated to a partition (hereinafter referred to as “colocate partition”) that occupies the same position as the target partition. It sets to temporal prediction vector candidate # 313.
- (B) of FIG. 3 is a diagram showing a target partition, a prediction vector assigned to the target partition, a co-located partition, and a motion vector assigned to the co-located partition.
- the prediction vector candidate selection unit 314 internally estimates the prediction accuracy of the temporal prediction vector candidate, and based on the estimation result, from the spatial prediction vector candidate and the temporal prediction vector candidate, a second prediction vector candidate # 314 The motion vector candidate used as is selected.
- any one of the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 is selected and set as the second prediction vector candidate # 314.
- the prediction vector candidate selection unit 314 selects any one of the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 according to the size of the target partition. It is selected and set as a second predicted vector candidate # 314.
- the predicted vector candidate selection unit 314 selects the temporally predicted vector candidate # 313, and the size of the target partition is greater than the above reference size. When it is smaller, the second spatial prediction vector candidate # 312 is selected.
- the size of the target partition can also be represented by using the number of pixels included in the target partition as an index, or the sum of the lengths of two adjacent sides of the target partition can also be represented as an index.
- the target partition may have a length of one of the two adjacent sides (for example, the larger one).
- the size of the target partition can also be represented by an index other than the above index.
- the motion vector assigned to the co-located partition is similar to the motion vector allocated to the target partition. That is, when the size of the target partition is large, the prediction accuracy of the motion vector allocated to the co-located partition is high.
- the motion vector assigned to the co-located partition is not similar to the motion vector assigned to the target partition. That is, when the size of the target partition is small, the prediction accuracy of the motion vector allocated to the co-located partition is low.
- the predicted vector candidate selection unit 314 selects the temporally predicted vector candidate # 313, and the size of the target partition is smaller than the above reference size
- the second spatial prediction vector candidate # 312 it is possible to select prediction vector candidates more similar to the motion vector assigned to the target partition, so that it is possible to improve the coding efficiency.
- the temporal prediction vector calculation unit 313 sets the median of the motion vector allocated to the co-located partition and the motion vector allocated to each of the partitions around the co-located partition to the temporal prediction vector candidate # 313. It may be set. More specifically, the temporal prediction vector calculation unit 313 calculates a median of the motion vector allocated to the co-located partition and the motion vector allocated to each of the partitions adjacent to the co-located partition as temporal prediction vector candidates. It may be set to # 313. In addition, the temporal prediction vector calculation unit 313 further includes each of the motion vector allocated to the co-located partition, the motion vector allocated to each of the partitions adjacent to the co-located partition, and each of the partitions sharing the vertex with the co-located partition. The median of the motion vector allocated to may be set as temporal prediction vector candidate # 313.
- the prediction vector candidate generation unit 31 is configured to calculate the first spatial prediction vector and the second prediction vector candidate # 314, the other motion vector candidates including the second prediction vector candidate # 314 are It does not matter as a structure to calculate. For example, a global motion vector to be assigned to each frame as another prediction vector candidate may be calculated as a motion vector candidate.
- the coding device since calculation of the second predicted vector candidate # 314 may increase the operation load of the encoding apparatus, information indicating whether to use the second predicted vector candidate # 314 is encoded in the encoded data It is also preferable that the coding device be switchable.
- the position in the coded data may be any of a sequence header, a frame header, a slice header, and a macroblock header, and further, coded data for controlling the configuration of coded data of a sequence and a picture, H .
- encoding may be performed using a sequence parameter set and a picture parameter set.
- FIG. 4 is a flowchart showing a first example of the operation of the motion vector redundancy reduction unit 19.
- the first spatial prediction vector calculation unit 311, the second spatial prediction vector calculation unit 312, and the temporal prediction vector calculation unit 313 are respectively the first spatial prediction vector candidate # 311 and the second space.
- a target predictive vector candidate # 312 and a temporally predicted vector candidate # 313 are calculated (step S1).
- the predicted vector candidate selection unit 314 sets the second spatial predicted vector candidate # 312 to the second predicted vector candidate # 314, and the size of the target partition Is equal to or larger than the reference size, the temporal prediction vector candidate # 313 is set as the second prediction vector candidate # 314 (step S2).
- the prediction vector selection unit 32 calculates a coding cost for each of the first spatial prediction vector candidate # 311 and the second prediction vector candidate # 314 (step S3).
- the prediction vector selection unit 32 sets, as a prediction vector # 32a, a motion vector candidate having a smaller coding cost among the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314. (Step S4).
- the PMV flag generation unit 33 sets a flag # 19b indicating which one of the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314 is set to the prediction vector # 32a. Generate (step S5).
- the flag # 19b is set to the prediction vector # 32a among the first spatial prediction vector candidate # 311, the second spatial prediction vector candidate # 312, or the second prediction vector candidate # 314. It may be a flag indicating whether it has been done.
- the subtractor 34 generates a differential motion vector # 19a which is a difference vector between the prediction vector # 32a and the motion vector # 17 allocated to the target partition (step S6).
- step S11 to step S19 in the motion vector redundancy reduction unit 19 are repeated for each target partition.
- the motion vector redundancy reduction unit 19 selects any one of the second spatial prediction vector candidate or the temporal prediction vector candidate according to the size of the target partition. By selecting according to the size of the target partition, it is possible to select, among the second spatial prediction vector candidates or temporal prediction vector candidates, prediction vector candidates that are more similar to the motion vector allocated to the target partition .
- Coding efficiency can be further improved by using prediction vector candidates that are more similar to the motion vector generally assigned to the target partition.
- the motion vector redundancy reduction unit 19 sets the selected prediction vector candidate as the second prediction vector candidate.
- the motion vector redundancy reduction unit 19 sets a prediction vector candidate having a small coding cost among the second prediction vector candidate or the first spatial prediction vector candidate as a prediction vector to be allocated to a target partition.
- coding efficiency can be further improved by using a prediction vector with lower coding cost.
- the candidate can be set to a motion vector to be assigned to the target partition.
- the prediction vector candidate selection unit 314 sets any one of the second spatial prediction vector candidate or the temporal prediction vector candidate as a second prediction vector candidate according to the size of the target partition. In order to select, the flag indicating which prediction vector candidate has been selected as the second prediction vector candidate is unnecessary.
- the decoding device that decodes the image based on the encoded data # 2 generated by the moving image encoding device 1 is the second one corresponding to the size of the target partition, as in the motion vector redundancy reduction unit 19. Since any one of the spatial prediction vector candidate or the temporal prediction vector candidate can be selected as the second prediction vector candidate according to the size of the target partition, the flag is referred to. Instead, it can be reproduced which prediction vector candidate has been selected as the second prediction vector candidate.
- the motion vector redundancy reduction unit 19 not only can the coding efficiency be further improved by using a prediction vector with a smaller coding cost, but any prediction vector candidate becomes a prediction vector. Since the code amount of the flag indicating whether it is selected can be reduced, the coding efficiency can be further enhanced.
- FIG. 5 is a flowchart showing a second example of the operation of the motion vector redundancy reduction unit 19.
- the first spatial prediction vector calculation unit 311 calculates a first spatial prediction vector candidate # 311 (step S11).
- the motion vector redundancy reduction unit 19 determines whether the skip mode is applied to the target partition (step S12).
- the prediction vector candidate selection unit 314 determines whether the size of the target partition is equal to or larger than a predetermined reference size (step S12). Step S13).
- the prediction vector candidate selection unit 314 causes the temporal prediction vector calculation unit 313 to calculate the temporal prediction vector candidate # 313.
- the temporal prediction vector candidate # 313 is set as a second prediction vector candidate # 314 (step S14).
- the prediction vector candidate selection unit 314 calculates the second spatial prediction vector candidate # 312, and sets the second spatial prediction vector candidate # 312 as the second prediction vector candidate # 314 ( Step S15).
- the prediction vector selection unit 32 calculates a coding cost for each of the first spatial prediction vector candidate # 311 and the second prediction vector candidate # 314 (step S16).
- the prediction vector selection unit 32 sets, as a prediction vector # 32a, a motion vector candidate having a smaller coding cost among the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314. (Step S17).
- the PMV flag generation unit 33 sets a flag # 19b indicating which one of the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314 is set to the prediction vector # 32a. It generates (step S18).
- the flag # 19b is set to the prediction vector # 32a among the first spatial prediction vector candidate # 311, the second spatial prediction vector candidate # 312, or the second prediction vector candidate # 314. It may be a flag indicating whether it has been done.
- the subtractor 34 generates a differential motion vector # 19a which is a difference vector between the prediction vector # 32a and the motion vector # 17 allocated to the target partition (step S19).
- step S11 to step S19 in the motion vector redundancy reduction unit 19 are repeated for each target partition.
- the motion vector redundancy reduction unit 19 calculates the temporal prediction vector candidate # 313, and the temporal prediction vector candidate # 313 is If the size of the target partition is smaller than the reference size, the second spatial prediction vector candidate # 312 is calculated, and the second spatial prediction vector candidate # 312 is calculated. Is set as the second predicted vector candidate # 314.
- the motion vector redundancy reduction unit 19 does not calculate the second spatial prediction vector candidate # 312, and the size of the target partition. Is smaller than the reference size, the temporally predicted vector candidate # 313 is not calculated.
- the motion vector redundancy reduction unit 19 determines whether the size of the target partition is equal to or greater than a predetermined reference size. Instead, the second spatial prediction vector candidate # 312 is set as a second prediction vector candidate # 314.
- the motion vector redundancy reduction unit 19 can not only calculate a prediction vector with high coding efficiency, but also calculates the calculation cost for calculating such a prediction vector. It can be minimized. Therefore, the motion vector redundancy reduction unit 19 can not only calculate a prediction vector high in coding efficiency, but also can calculate such a prediction vector at high speed.
- FIG. 6 is a block diagram showing the configuration of the moving picture decoding apparatus 2 according to the present invention.
- the moving picture decoding apparatus 2 includes a variable-length code decoding unit 23, a motion vector restoring unit 24, a buffer memory 25, a predicted image generating unit 26, an intra predicted image generating unit 27, and a prediction scheme determining unit 28.
- An inverse quantization / inverse transform unit 29 and an adder 30 are provided.
- the video decoding device 2 receives the encoded data # 2 and sequentially outputs an output image # 3.
- variable-length code decoding unit 23 performs variable-length decoding on the encoded data # 2, and outputs a differential motion vector # 23a, a prediction mode # 23b, a quantized prediction residual data # 23c, and a PMV flag 23d.
- the motion vector restoration unit 24 decodes the motion vector # 24 of the target macroblock from the motion vector # 25a that has already been decoded and is stored in the buffer memory 25.
- the motion vector restoration unit 24 will not be described here because it will be described in detail below.
- the buffer memory 25 stores a later-described decoded image # 3, motion vector # 24, and prediction mode # 23b.
- the predicted image generation unit 26 generates an inter predicted image # 26 from the motion vector # 25a and the decoded image # 3 stored in the buffer memory 25.
- the intra predicted image generation unit 27 generates an intra predicted image # 27 from the locally decoded image # 25 b in the same image as the target macroblock stored in the buffer memory 25.
- the prediction method determination unit 28 selects one of the intra prediction image # 27 and the inter prediction image # 26 based on the prediction mode # 23 b, and outputs it as a prediction image # 28.
- the inverse quantization / inverse transform unit 29 performs inverse quantization and inverse DCT transform on the quantized prediction residual data # 23c, and outputs a prediction residual # 29.
- the adder 30 adds the quantized prediction residual data # 23c and the predicted image # 28 and outputs the result as a decoded image # 3. Also, the output decoded image # 3 is stored in the buffer memory 25.
- FIG. 7 is a block diagram showing the configuration of the motion vector restoration unit 24. As shown in FIG.
- the motion vector restoration unit 24 includes a prediction vector candidate generation unit 31, a prediction vector determination unit 35, and an adder 36.
- the prediction vector candidate generation unit 31 receives the motion vector # 25a accumulated in the buffer memory 25 and outputs a first spatial prediction vector candidate # 311 and a second prediction vector candidate # 314 for the target partition. .
- the specific configuration and operation of the predicted vector candidate generation unit 31 have already been described, so the description will be omitted here.
- the prediction vector determination unit 35 selects one of the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314 based on the PMV flag # 23d output from the variable-length code decoding unit 23. One prediction vector candidate is selected as the prediction vector # 35 to be allocated to the target partition. The selected prediction vector # 35 is output to the adder.
- the PMV flag # 23d is the same flag as the PMV flag # 19b generated in the video encoding device 1.
- variable length code decoding unit 23 decodes the PMV flag # 23 d in order to clarify the correspondence with the moving picture encoding device 1, either of the direct The selection information (described as # 32b in the moving picture coding apparatus), which is information indicating whether or not the predicted vector candidate is selected, may be directly decoded.
- the prediction vector # 35 selected by the prediction vector determination unit 35 is predicted by the video encoding device 1 This is a prediction vector identical to the prediction vector # 32a selected by the vector selection unit 32.
- the adder 36 adds the difference motion vector # 23a output from the variable-length code decoding unit 23 and the prediction vector # 35 output from the prediction vector determination unit 35, and restores the motion vector # 24.
- FIG. 8 is a flowchart showing the operation of the motion vector restoration unit 24.
- the motion vector restoration unit 24 analyzes information indicating whether or not the skip mode is applied, which is embedded in the block mode information of the encoded data # 2 (step S21).
- the first spatial prediction vector calculation unit 311 calculates a first spatial prediction vector candidate # 311 (step S22).
- the motion vector restoration unit 24 determines whether the skip mode is applied to the target partition (step S23).
- the prediction vector candidate selection unit 314 determines whether the size of the target partition is equal to or greater than a predetermined reference size (step S23). Step S24).
- the prediction vector candidate selection unit 314 causes the temporal prediction vector calculation unit 313 to calculate the temporal prediction vector candidate # 313.
- the temporal prediction vector candidate # 313 is set as the second prediction vector candidate # 314 (step S26).
- the prediction vector candidate selection unit 314 calculates the second spatial prediction vector candidate # 312, and sets the second spatial prediction vector candidate # 312 as the second prediction vector candidate # 314 ( Step S25).
- the prediction vector determination unit 35 analyzes the PMV flag # 23 d (step 27), and among the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 314, the PMV flag # 23 d Is selected and set as a prediction vector # 35 (step S28).
- the prediction vector # 35 is output to the adder # 36.
- the adder 36 adds the prediction vector # 35 output from the prediction vector determination unit 35 and the differential motion vector # 23a output from the variable length code decoding unit 23 to generate a motion vector # 24. (Step 29).
- steps S21 to S29 in the motion vector restoration unit 24 are repeated for each target partition.
- the prediction vector # 35 can be generated.
- the generated predicted vector # 35 is the same vector as the predicted vector # 32a used in the encoding process in the video encoding device 1. Therefore, the motion vector # 24 generated by the motion vector restoration unit 24 is the same motion vector as the motion vector # 17 generated in the video encoding device 1.
- the motion vector restoration unit 24 it is possible to restore the same motion vector # 24 as the motion vector # 17 generated in the video encoding device 1.
- the output picture # 3 can be generated based on the coded data # 2 with high coding efficiency.
- the prediction vector candidate selection unit 314 sets any one of the second spatial prediction vector candidate or the temporal prediction vector candidate as a second prediction vector candidate according to the size of the target partition. For selection, the second predicted vector can be restored without referring to a flag indicating which predicted vector candidate has been selected as the second predicted vector candidate.
- the moving picture decoding apparatus 2 that decodes an image based on the encoded data # 2 generated by the moving picture encoding apparatus 1 is a second spatial prediction vector candidate or Since any one of temporal prediction vector candidates can be selected as a second prediction vector candidate according to the size of the target partition, any prediction vector candidate can be selected without referring to the flag. It can be reproduced whether it is selected as a second order predicted vector candidate.
- the code amount of the flag indicating which prediction vector candidate is selected as the prediction vector can be reduced, the coding efficiency can be further enhanced.
- FIG. 9 is a diagram showing a bit stream #MB for each macroblock of the encoded data # 2.
- N represents the number of partitions included in a macroblock.
- the flag # 19b is included in the bitstream #MB only when it is necessary to select a prediction vector of each partition.
- the block mode information Mod includes the prediction mode # 18 b of the macro block, partition division information, and the like. Further, the block mode information Mod includes information indicating whether or not the skip mode is applied to each partition.
- the index information Idxi includes a reference picture number referred to by each partition, which is required when performing motion compensation.
- the motion vector information MVi includes a differential motion vector # 19a for each partition.
- the moving picture decoding apparatus 2 can restore the same prediction vector # 24 as the prediction vector # 19a based on the encoded data # 2 configured by the bit stream #MB. Also, the moving picture decoding apparatus 2 can generate an output image # 3 based on such a predicted vector # 24.
- bit stream #MB has the above-described structure, it is possible to realize the coded data # 2 that can be decoded by the moving picture decoding apparatus 2 and has high coding efficiency.
- FIG. 10 is a block diagram showing a configuration of the motion vector redundancy reduction unit 19 included in the moving picture decoding apparatus according to the present modification.
- the other configuration of the moving picture coding apparatus according to the present modification is the same as that of the moving picture coding apparatus 1 according to the first embodiment.
- the prediction vector candidate generation unit 31 includes a prediction vector candidate selection unit 315 instead of the prediction vector candidate selection unit 314.
- the prediction vector candidate selection unit 315 selects any one of the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 according to the length of the motion vector allocated to the target partition. A prediction vector candidate is selected and set as a second prediction vector candidate # 315.
- the prediction vector candidate selection unit 315 selects the temporal prediction vector candidate # 313, and selects the target partition.
- the second spatial prediction vector candidate # 312 is selected.
- the second predicted vector candidate # 315 is output to the predicted vector selection unit 32.
- the prediction vector selection unit 32 selects any one of the first spatial prediction vector candidate # 311 or the second prediction vector candidate # 315, which has high coding efficiency, Set to prediction vector # 32a.
- the motion of the image represented by the target partition is large. In such a case, the correlation between the image represented by the co-located partition and the image represented by the target partition is low.
- the motion vector allocated to the co-located partition is not similar to the motion vector allocated to the target partition. In other words, when the length of the motion vector allocated to the target partition is large, the prediction accuracy of the motion vector allocated to the co-located partition is low.
- the motion vector allocated to the target partition when the motion vector allocated to the target partition is short, the correlation between the image represented by the co-located partition and the image represented by the target partition is high, and the motion vector allocated to the co-located partition is Similar to the motion vector assigned to the target partition. In other words, when the length of the motion vector allocated to the target partition is short, the prediction accuracy of the motion vector allocated to the co-located partition is high.
- the prediction vector candidate selection unit 315 selects the temporal prediction vector candidate # 313 and sets it as the target partition.
- the second spatial prediction vector candidate # 312 can be selected, so that the coding efficiency can be improved.
- the motion vector redundancy reduction unit 19 in the present modification performs substantially the same operation as the motion vector redundancy reduction unit 19 in the first embodiment except for the above point.
- the moving picture coding apparatus according to the modification of the present invention includes a predicted vector candidate selection section 315 ′ in place of the predicted vector candidate selection section 314 in the moving picture coding apparatus 1.
- the other configuration of the moving picture coding apparatus according to the present modification is the same as that of the moving picture coding apparatus 1 according to the first embodiment.
- the prediction vector candidate selection unit 315 ′ selects one of the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 according to the length of the spatial prediction vector allocated to the target partition. One prediction vector candidate is selected and set as a secondary prediction vector candidate # 315.
- the prediction vector candidate selection unit 315 ′ selects the temporal prediction vector candidate # 313. If the length of the spatial prediction vector assigned to the target partition is larger than a predetermined reference interval, the second spatial prediction vector candidate # 312 is selected.
- the motion of the image represented by the target partition is large. In such a case, the correlation between the image represented by the co-located partition and the image represented by the target partition is low.
- the motion vector allocated to the co-located partition is not similar to the motion vector allocated to the target partition. In other words, when the length of the spatial prediction vector allocated to the target partition is large, the prediction accuracy of the motion vector allocated to the co-located partition is low.
- the spatial prediction vector allocated to the target partition is short, the correlation between the image represented by the co-located partition and the image represented by the target partition is high, and the motion vector allocated to the co-located partition is Is similar to the motion vector assigned to the target partition. In other words, when the length of the spatial prediction vector assigned to the target partition is short, the prediction accuracy of the motion vector assigned to the co-located partition is high.
- the prediction vector candidate selection unit 315 ′ selects the temporal prediction vector candidate # 313, Since the second spatial prediction vector candidate # 312 can be selected when the length of the spatial prediction vector allocated to the target partition is larger than a predetermined reference value, the coding efficiency can be improved. it can.
- the motion vector redundancy reduction unit 19 in the present modification performs substantially the same operation as the motion vector redundancy reduction unit 19 in the first embodiment except for the above point.
- the prediction vector candidate selection unit 315 ′ does not depend on the length of the spatial prediction vector allocated to the target partition, but instead on the second spatial prediction vector according to the length of the temporal prediction vector candidate # 313. It is also possible to select any one prediction vector candidate from among the candidate # 312 or the temporal prediction vector candidate # 313 and set it as the secondary prediction vector candidate # 315.
- the moving picture coding apparatus according to the second modified example of the present invention includes a predicted vector candidate selection section 316 instead of the predicted vector candidate selection section 314 in the moving picture coding apparatus 1.
- the other configuration of the moving picture coding apparatus according to the present modification is the same as that of the moving picture coding apparatus 1 according to the first embodiment.
- the prediction vector candidate selection unit 316 selects either the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 according to the frame interval between the frame including the target partition and the frame including the co-located partition. One of the prediction vector candidates is selected and set as a second prediction vector candidate # 316.
- the prediction vector candidate selection unit 316 determines that the temporal prediction vector candidate # If the frame interval between the frame including the target partition and the frame including the co-located partition is larger than a predetermined reference interval, the second spatial prediction vector candidate # 312 is selected.
- the frame interval of each frame can be detected by referring to the relative position of the reference image given to each frame.
- (A) of FIG. 11 is a figure which shows the reference image relative position provided to the flame
- FIG. 11 is a diagram illustrating a target frame and an image number assigned to a frame near the target frame.
- the correlation between the image represented by the target partition and the image represented by the co-located partition is low.
- the motion vector allocated to the co-located partition is not similar to the motion vector allocated to the target partition. In other words, when the frame interval between the frame including the target partition and the frame including the co-located partition is long, the prediction accuracy of the motion vector allocated to the co-located partition is low.
- the motion vector allocated to the co-located partition is similar to the motion vector allocated to the target partition.
- the prediction accuracy of the motion vector allocated to the co-located partition is high.
- the prediction vector candidate selection unit 316 selects the temporally predicted vector candidate # 313. And select the second spatial prediction vector candidate # 312 if the frame interval between the frame including the target partition and the frame including the co-located partition is greater than a predetermined reference interval. Since it is possible, it is possible to select a second prediction vector candidate with high coding efficiency.
- the motion vector redundancy reduction unit 19 in the present modification performs substantially the same operation as the motion vector redundancy reduction unit 19 in the first embodiment except for the above point.
- the moving picture decoding apparatus includes the above-described predicted vector candidate selection section 316 instead of the predicted vector candidate selection section 314 in the moving picture decoding apparatus 2 according to the first embodiment. That is, when the moving picture decoding apparatus according to the present modification determines that the frame interval between the frame including the target partition and the frame including the co-located partition is equal to or less than a predetermined reference interval, the temporal prediction vector candidate # 313 Is selected, and if the frame interval between the frame including the target partition and the frame including the co-located partition is larger than a predetermined reference interval, the second spatial prediction vector candidate # 312 is selected.
- the moving picture coding apparatus according to the third modification of the present invention is provided with a predicted motion vector candidate selecting section 317 instead of the predicted motion vector candidate selecting section 314 in the moving picture coding apparatus 1.
- the other configuration of the moving picture coding apparatus according to the present modification is the same as that of the moving picture coding apparatus 1 according to the first embodiment.
- the predicted vector candidate selection unit 317 has a GOP (Group of Picture) structure in which one or both of the frame including the target partition and the frame including the co-located partition are configured by a picture including B picture. If it belongs to (Picture group structure), the second spatial prediction vector candidate # 312 is selected, and a frame containing the target partition and a frame containing the co-located partition do not belong to the above GOP structure. , Temporally-prediction vector candidate # 313 is selected.
- GOP Group of Picture
- any one of the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 is selected and set as a secondary prediction vector candidate # 317.
- FIG. 12 is a diagram showing a GOP structure configured by, in particular, layer B pictures among GOP structures including B pictures. As shown in FIG. 12, the GOP structure is composed of I picture (Intra picture), P picture (Predictive picture), and B picture (Bi-directional predictive picture).
- the frame interval tends to be long. Therefore, if any one or both of the frame including the target partition and the frame including the co-located partition belong to the above GOP structure, the temporal prediction vector candidate # 313 is allocated to the target partition It is likely not to be similar to the motion vector being That is, when one or both of the frame including the target partition and the frame including the co-located partition belong to the above GOP structure, the prediction accuracy of the temporal prediction vector candidate # 313 is low.
- the moving picture coding apparatus of the present modification when one or both of the frame including the target partition and the frame including the co-located partition belong to the above GOP structure, 2 spatial prediction vector candidate # 312 is selected, and if the frame including the target partition and the frame including the co-located partition do not belong to the above GOP structure, temporal prediction vector candidate # 313 is selected Therefore, it is possible to select a second prediction vector candidate with high coding efficiency.
- the motion vector redundancy reduction unit 19 in the present modification performs substantially the same operation as the motion vector redundancy reduction unit 19 in the first embodiment except for the above point.
- coding in which the above GOP structure is applied to one or both of the frame including the target partition and the frame including the co-located partition is used. It is preferable to generate a flag #BC as to whether or not it has been sent and send it to the decoding device.
- the moving picture decoding apparatus also includes a predicted vector candidate selection unit 317 instead of the predicted vector candidate selection unit 314 in the moving picture decoding apparatus 2. That is, also in the moving picture decoding apparatus according to the present modification, when one or both of the frame including the target partition and the frame including the co-located partition belong to the above GOP structure, As the spatial prediction vector candidate # 312 is selected, and the frame including the target partition and the frame including the co-located partition do not belong to the above GOP structure, the temporal prediction vector candidate # 313 is selected. , It is possible to select a second prediction vector candidate with high coding efficiency. Note that the moving picture decoding apparatus according to the present modification can determine whether or not the GOP structure is applied by referring to the flag #BC.
- coded data with high coding efficiency can be decoded by using the moving picture decoding apparatus according to the present modification.
- a moving picture coding apparatus according to a third modified example of the present invention will be described.
- the moving picture coding apparatus according to the third modified example of the present invention is provided with a predicted motion vector candidate selection section 317 ′ instead of the predicted motion vector candidate selection section 314 in the moving picture coding apparatus 1.
- the other configuration of the moving picture coding apparatus according to the present modification is the same as that of the moving picture coding apparatus 1 according to the first embodiment.
- a moving picture coding apparatus according to a third modified example of the present invention will be described.
- the prediction vector candidate selection unit 317 ′ selects the second spatial prediction vector candidate # 312, and the frame containing the target partition is not a B picture. , Select the temporal prediction vector candidate # 313.
- the frame interval tends to be wide. In this modification, it is possible to use this tendency without encoding the flag #BC as to whether or not the GOP structure is applied.
- the motion vector of the target partition is more similar to the motion vector of the target partition by determining the prediction accuracy of the temporal prediction motion vector # 313 and changing the selection to the second prediction vector candidate # 314 according to the type of frame. It is possible to use motion vector candidates that are highly likely to
- the moving picture decoding apparatus according to the present modification is also the same as the moving picture decoding apparatus 2 according to the first embodiment except that a prediction vector candidate selection unit 317 'is provided instead of the prediction vector candidate selection unit 314. .
- coded data with high coding efficiency can be decoded by using the moving picture decoding apparatus according to the present modification.
- the moving picture coding apparatus according to the third ′ ′ modification of the present invention includes a prediction vector candidate selection section 317 ′ ′ instead of the prediction vector candidate selection section 314 in the moving picture coding apparatus 1.
- the other configuration of the moving picture coding apparatus according to the present modification is the same as that of the moving picture coding apparatus 1 according to the first embodiment.
- a moving picture coding apparatus according to a third variation of the present invention will be described.
- the predicted vector candidate selection unit 317 selects the second spatial predicted vector candidate # when the frame including the target partition is a B picture and the size of the target partition is equal to or larger than a predetermined reference size. In step S 312, the temporal prediction vector candidate # 313 is selected.
- the predicted vector candidate selection unit 317 determines whether the frame including the target partition is a B picture or the size of the target partition is equal to or greater than a predetermined reference size in the second spatial
- the prediction vector candidate # 312 may be selected, and in other cases, the temporal prediction vector candidate # 313 may be selected.
- the prediction accuracy of the temporal motion vector predictor # 313 is determined according to the type of the frame and the size of the target partition, and the selection of the second predicted vector candidate # 314 is changed to further increase the target partition. It is possible to use motion vector candidates that are likely to be similar to motion vectors of.
- the moving picture decoding apparatus is the same as the moving picture decoding apparatus 2 according to the first embodiment except that a prediction vector candidate selection unit 317 '' is provided instead of the prediction vector candidate selection unit 314. is there.
- coded data with high coding efficiency can be decoded by using the moving picture decoding apparatus according to the present modification.
- the moving picture coding apparatus according to the fourth modification of the present invention includes a predicted vector candidate selection section 318 in place of the predicted vector candidate selection section 314 in the moving picture coding apparatus 1.
- the other configuration of the moving picture coding apparatus according to the present modification is the same as that of the moving picture coding apparatus 1 according to the first embodiment.
- the prediction vector candidate selection unit 318 selects the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 according to the size of the target partition and the length of the motion vector allocated to the target partition. One of the prediction vector candidates is selected and set as a second prediction vector candidate # 318.
- the predicted vector candidate selection unit 318 sets the length of the motion vector allocated to the target partition to a predetermined reference. If it is equal to or less than the value, the temporal prediction vector candidate # 313 is selected, and otherwise, the second spatial prediction vector candidate # 312 is selected.
- the temporal prediction vector # 313 is similar to the motion vector allocated to the target partition. Also, even when the length of the motion vector allocated to the target partition is small, the temporal prediction vector # 313 is similar to the motion vector allocated to the target partition. In other words, when the size of the target partition is large, the prediction accuracy of the temporal prediction vector # 313 is high, and even when the size of the target partition is small, the prediction accuracy of the temporal prediction vector # 313 is high.
- the motion vector redundancy reduction unit 19 in the present modification performs substantially the same operation as the motion vector redundancy reduction unit 19 in the first embodiment except for the above point.
- the moving picture decoding apparatus also includes the above-described predicted vector candidate selection unit 318 instead of the predicted vector candidate selection unit 314 in the moving picture decoding apparatus 2 according to the first embodiment. That is, the moving picture decoding apparatus according to the present modification is also a case where the size of the target partition is equal to or larger than a predetermined reference size, and the length of the motion vector allocated to the target partition is predetermined. If it is equal to or less than the value, the temporal prediction vector candidate # 313 is selected, and otherwise, the second spatial prediction vector candidate # 312 is selected.
- the prediction vector candidate selection unit 318 selects the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate according to the size of the target partition or the length of the motion vector allocated to the target partition. It is also possible to select any one prediction vector candidate from # 313 and set it as the second prediction vector candidate # 315.
- the predicted vector candidate selection unit 315 determines the length of the motion vector allocated to the target partition in advance.
- the temporal prediction vector candidate # 313 may be selected, and in the other cases, the second spatial prediction vector candidate # 312 may be selected.
- the prediction vector candidate selection unit 318 selects the length of the first spatial prediction vector candidate # 311 and the second spatial prediction vector candidate # 312. According to the length of L or the length of temporal prediction vector candidate # 313, any one of the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313 May be selected.
- the predicted vector candidate selection unit 318 selects the temporally predicted vector candidate # 313, and selects the target partition. If the length of the motion vector allocated to is larger than the side length of the target partition, the second spatial prediction vector candidate # 312 may be selected.
- the prediction vector candidate selection unit 314 (also 315, 316, 317, 318, etc.) is one of the second spatial prediction vector candidate # 312 or the temporal prediction vector candidate # 313. Although one is selected, both may be selected. More specifically, when it is determined in the prediction vector candidate selection unit 314 that the prediction accuracy of the temporal prediction vector candidate # 313 of the target partition is high, the second spatial prediction vector candidate # 312 and time are determined. In the case where both of the target predictive vector candidate # 313 are selected and it is determined that the prediction accuracy of the temporal predicted vector candidate # 313 of the target partition is low, the second spatial predicted vector candidate # 312 is selected.
- both the temporal prediction vector candidate and the spatial prediction vector candidate are selected, and the size of the target partition is smaller than the reference size
- the temporal prediction vector candidate # 313 is selected when it is determined that the prediction accuracy of the temporal prediction vector candidate # 313 is high, particularly because the spatial correlation of motion is large in the skip block. Not only that, but it is preferable to simultaneously select the second spatial prediction vector candidate # 312.
- the prediction of the temporal prediction vector is based on the frame interval, the type and structure of the frame, the length of the motion vector, and both the length of the motion vector and the size of the target partition.
- a method of estimating and determining the accuracy is also preferable.
- the first spatial prediction vector # 311 and the second prediction vector candidate # 314 are selected in the coding device by the prediction vector selection unit 32 and the decoding device in the prediction vector determination unit 35.
- the second prediction vector candidate # 314 (which may be plural as shown in the supplementary item 1) is as it is, the prediction vector selection unit 32 and the prediction vector determination It is also possible to use as a prediction vector candidate used in the part 35.
- the flag # 19b indicating which prediction vector was used to generate the differential motion vector # 19a is unnecessary, and the number of secondary prediction vector candidates is In the case where there are plural, the flag # 19b is required.
- the second spatial prediction vector # 312 used for calculation of the second prediction vector candidate # 314 is the median of the motion vector allocated to each of the left partition, upper partition, and upper right partition of the target partition It is preferable to use
- the temporal prediction vector # 313 is used as a prediction vector candidate when the prediction accuracy of the temporal prediction vector # 313 is high, and otherwise the spatial prediction vector # 312 (
- the above-mentioned median is used as a prediction vector candidate. It is said that.
- the temporal prediction vector candidate # 313 should be added to the option even in the skip block. Is considered suitable.
- the size of the target partition is large, a large improvement in image quality can be expected by increasing the choice of prediction vectors as compared to the case where the size is small, so that temporal prediction vector candidate # 313 may be added in the skip block. It is suitable.
- motion vector candidates for which the size of the target partition is equal to or larger than the reference size are the first spatial prediction vector candidate # 311 and the second spatial prediction vector candidate # 312.
- the temporal vector predictor candidate # 313 is a motion vector candidate when the size of the target partition is smaller than the reference size, the first spatial vector predictor candidate # 311 and the second spatial vector predictor candidate # 312 .
- motion vector candidates when the size of the target partition is equal to or larger than the reference size are the first spatial prediction vector candidate # 311 and the temporal prediction vector candidate # 313, and the size of the target partition is the reference size
- the motion vector candidates in the case of being smaller than the first spatial prediction vector candidate # 311 and the second spatial prediction vector candidate # 312 are used.
- the determination based on the size of the target partition is to determine the prediction accuracy of the temporally-predicted vector candidate # 313, and another determination such as a frame interval may be used.
- the moving picture coding apparatus includes a median of motion vectors allocated to each of a left partition adjacent to the left side of the target partition, an upper partition adjacent to the upper side of the target partition, and an upper right partition adjacent to the right side of the upper partition.
- a first calculation means for setting the first prediction vector candidate as a candidate of a prediction vector to be allocated to the target partition, and a candidate of a prediction vector to allocate a motion vector allocated to the left partition to the object partition A second calculation means for setting 2 predicted vector candidates, and a predicted vector for allocating a motion vector allocated to the co-located partition occupying the same position as the target partition in the encoded frame to the target partition.
- the image processing apparatus further comprises second selecting means for selecting any one of the prediction vector candidates and setting it as the prediction vector to be allocated to the target partition.
- any one of the second prediction vector candidate or the third prediction vector candidate is selected. Since the first selecting means is provided, any one of the second predicted vector candidate and the third predicted vector candidate is selected according to the prediction accuracy of the third predicted vector candidate. Predictive vector candidates can be selected.
- the first prediction vector candidate or the prediction vector candidate selected by the first selection means may be selected according to the coding cost.
- the present invention includes the second selecting means for selecting any one predicted vector candidate from among the above and setting the selected predicted vector as the predicted vector to be allocated to the target partition.
- the second selection means selects a prediction vector candidate having a smaller coding cost among the first prediction vector candidate or the prediction vector candidate selected by the first selection means, It is preferable to set to the prediction vector allocated to the said object partition.
- the second selection unit is a prediction vector whose coding cost is smaller among the first prediction vector candidate or the prediction vector candidate selected by the first selection unit. Since a candidate can be selected and set as a prediction vector to be assigned to the target partition, it is possible to generate a prediction vector with high coding efficiency, which has a further effect of generating a prediction vector.
- the first selecting means selects the third predicted vector candidate when the size of the target partition is equal to or larger than a predetermined reference size, and when the size of the target partition is smaller than the reference size.
- the second prediction vector candidate is selected.
- the prediction accuracy of the third predicted vector is lower, and when the size of the target partition is larger, the prediction accuracy of the third predicted vector is higher. is there.
- the third predicted vector candidate when the size of the target partition is equal to or larger than a predetermined reference size, the third predicted vector candidate is selected, and when the size of the target partition is smaller than the reference size, the third predicted vector candidate is selected. Since two prediction vector candidates can be selected, prediction vector candidates with higher coding efficiency can be selected even when the size of the target partition is small or when the size of the target partition is large. The additional effect of being able to be selected is achieved.
- the first selection means selects the third predicted vector candidate and allocates the selected candidate to the target partition.
- the length of the motion vector is longer than the predetermined length, it is preferable to select the second prediction vector candidate.
- the prediction accuracy of the third prediction vector is higher, and the length of the motion vector allocated to the target partition is longer, the above There is a tendency that the prediction accuracy of the third prediction vector is lower.
- the third predicted vector candidate is selected and the motion allocated to the target partition
- the second predicted vector candidate can be selected, so even if the length of the motion vector allocated to the target partition is short, Even if it is long, there is a further effect that it is possible to select a prediction vector candidate with higher coding efficiency.
- the first selection means determines the third predicted vector when the frame interval between the frame including the target partition and the frame including the co-located partition is equal to or less than a predetermined reference frame interval.
- the candidate is selected, and when the frame interval between the frame including the target partition and the frame including the co-located partition is larger than the reference frame interval, the second predicted vector candidate is selected.
- the prediction accuracy of the third predicted vector candidate is higher, and the frame interval is longer. There is a tendency that the prediction accuracy of the third prediction vector candidate is lower.
- the third predicted vector candidate is selected.
- the second predicted vector candidate can be selected, so the target partition Further, it is possible to select a candidate for a prediction vector with higher coding efficiency even if the frame interval between the frame including the frame and the frame including the co-located partition is long or short. Play an effect.
- the first selection unit determines that at least one of the frame including the target partition or the frame including the co-located partition belongs to a picture group structure configured by a hierarchy B picture
- the second prediction vector candidate is selected, and when both the frame including the target partition and the frame including the co-located partition do not belong to the picture group structure configured by the layer B picture, the third prediction It is preferable to select vector candidates.
- the prediction accuracy of the third prediction vector candidate Tend decrease.
- the second prediction is performed when at least one of the frame including the target partition or the frame including the co-located partition belongs to the picture group structure configured by the layer B picture. If the vector candidate is selected and both the frame including the target partition and the frame including the co-located partition do not belong to the picture group structure configured by the layer B picture, the third predicted vector candidate is selected. Since it is possible to select, even if at least one of the frame including the target partition or the frame including the co-located partition belongs to the picture group structure configured by the layer B picture, the code A further effect of being able to choose a higher vector predictor candidate of efficiency.
- the first selecting means is a case where the size of the target partition is equal to or larger than a predetermined reference size, and the length of the motion vector allocated to the target partition is a predetermined length. In the following cases, it is preferable to select the third predicted vector candidate, and in other cases, to select the second predicted vector candidate.
- the prediction accuracy of the third predicted vector is lower, and when the size of the target partition is larger, the prediction accuracy of the third predicted vector is higher. is there. Also, in general, when the length of the motion vector allocated to the target partition is shorter, the prediction accuracy of the third prediction vector is higher, and the length of the motion vector allocated to the target partition is longer. There is a tendency that the prediction accuracy of the third prediction vector is lower.
- the size of the target partition is equal to or larger than the predetermined reference size, and the length of the motion vector allocated to the target partition is equal to or smaller than the predetermined length.
- the third predicted vector candidate is selected, and in the other cases, the second predicted vector candidate can be selected, Even when the size of the target partition is small, or even when the size of the target partition is large, it is possible to select a prediction vector candidate with higher coding efficiency, and it is allocated to the target partition. Even when the length of the motion vector is short or long, a prediction vector candidate with higher coding efficiency can be selected.
- the first selection means selects the third predicted vector candidate when the size of the target partition is equal to or greater than a predetermined reference size, and the target partition is selected. It is preferable to select the second predicted vector candidate when the size of is smaller than the reference size.
- the third predicted vector candidate when the size of the target partition is equal to or larger than a predetermined reference size, the third predicted vector candidate is selected, and when the size of the target partition is smaller than the reference size, the third predicted vector candidate is selected. Since it is possible to select two prediction vector candidates, it is possible to select a code without requiring a flag as to which one of the second prediction vector candidate and the third prediction vector candidate is to be selected. This has the further effect of being able to select a prediction vector candidate with higher conversion efficiency.
- the first selection means selects the third predicted vector candidate when the length of the motion vector allocated to the target partition is equal to or less than a predetermined length. It is preferable to select the second prediction vector candidate when the length of the motion vector allocated to the target partition is longer than the predetermined length.
- the third predicted vector candidate is selected and the motion allocated to the target partition
- the second predicted vector candidate can be selected, so that, among the second predicted vector candidate or the third predicted vector candidate, This has the further effect of being able to select a candidate for a prediction vector with higher coding efficiency without requiring a flag as to which to select.
- the first selection means determines that a frame interval between a frame including the target partition and a frame including the co-located partition is equal to or less than a predetermined reference frame interval.
- the third prediction vector candidate is selected and the frame interval between the frame including the target partition and the frame including the co-located partition is larger than the reference frame interval, the second prediction vector is selected. It is preferable to select a candidate.
- the third predicted vector candidate is selected.
- the second predicted vector candidate can be selected. Further, it is possible to select a candidate with higher coding efficiency without requiring a flag as to which of the above candidate predictor candidates or the above third candidate predictor candidate to select. Play an effect.
- the first selection means includes a flag indicating that the encoded image belongs to a picture group structure constituted by hierarchy B pictures in the encoded data. In the above case, it is preferable to select the second prediction vector candidate.
- the encoded data includes a flag indicating that the encoded image belongs to a picture group structure configured by layer B pictures
- the second predicted vector Since a candidate can be selected, even when the encoded image belongs to a picture group structure configured by layer B pictures, encoded data generated with high encoding efficiency can be decoded. It plays an effect.
- the first selection unit is a case where the size of the target partition is equal to or larger than a predetermined reference size, and the motion vector allocated to the target partition. If the length is equal to or less than a predetermined length, it is preferable to select the third predicted vector candidate, and in other cases, select the second predicted vector candidate.
- the size of the target partition is equal to or larger than the predetermined reference size, and the length of the motion vector allocated to the target partition is equal to or smaller than the predetermined length.
- the third prediction vector candidate can be selected, and in the other cases, the second prediction vector candidate can be selected, so that the second prediction vector candidate or the third prediction vector candidate can be selected.
- the prediction vector candidate having higher coding efficiency can be selected without requiring a flag as to which of the three prediction vector candidates is to be selected.
- the present invention can be suitably applied to a moving picture coding apparatus for coding moving pictures. Further, the present invention can be suitably applied to a moving picture decoding apparatus that restores a moving picture based on encoded data.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
対象フレームにおいて対象パーティション周辺の符号化済パーティションに割り付けられた動きベクトルからなる空間的動きベクトル群を参照し、算出する予測ベクトルを以下、空間的予測ベクトルと呼ぶ。上記空間的動きベクトル群を参照して空間的予測ベクトルを算出する方法の一つとして、対象パーティションの左辺に隣接する左パーティション、対象パーティションの上辺に隣接する上パーティション、および、当該上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルのメジアンをとる方法が挙げられる。
対象パーティションの左辺に隣接する左パーティションに割り付けられた動きベクトルを用いても良いし、
対象パーティションの上辺に隣接する上パーティションに割り付けられた動きベクトルを用いても良いし、
対象パーティションの左辺に隣接する左パーティション、対象パーティションの上辺に隣接する上パーティション、および、当該上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルの平均を用いても良い。
符号化済フレームにおいて対象パーティションと同じ位置を占めるパーティション(以下、コロケートパーティションと呼ぶ)、及び、コロケートパーティション周辺のパーティションに割り付けられた動きベクトルからなる時間的動きベクトル群を参照して、算出する、予測ベクトルを以下、時間的予測ベクトルと呼ぶ。上記時間的動きベクトル群を参照して時間的予測ベクトルを算出する方法の一つには、符号化済フレームにおいて対象パーティションと同じ位置を占めるコロケートパーティションの動きベクトルを用いる方法がある。時間的予測ベクトルを算出する方法としては、
例えば、
コロケートパーティションの左辺に隣接する左パーティション、コロケートパーティションの上辺に隣接する上パーティション、および、上記上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルのメジアンを用いても良いし、それらの平均を用いても良いし、
コロケートパーティションから、空間的予測ベクトル群のメジアン、もしくは、平均の動きベクトルの分、空間的に変位させた位置のパーティションに割り付けられた動きベクトルを用いても良い。
以下では、動きベクトル冗長性削減部の構成について、図1および図3の(a)~図3の(b)を参照して説明する。
以下では、予測ベクトル候補生成部31の構成について、図1および図3の(a)~図3の(b)を参照して説明する。図1に示すように、予測ベクトル候補生成部31は、第1空間的予測ベクトル算出部311、第2空間的予測ベクトル算出部312、時間的予測ベクトル算出部313、および、予測ベクトル候補選択部314を備えている。
以下では、動きベクトル冗長性削減部19の動作の第1の例について、図4を参照しつつ説明する。図4は、動きベクトル冗長性削減部19の動作の第1の例を示すフローチャートである。
動きベクトル冗長性削減部19の動作は、上記の動作例に限定されるものではない。以下では、動きベクトル冗長性削減部19の動作の第2の例について、図5を参照しつつ説明する。図5は、動きベクトル冗長性削減部19の動作の第2の例を示すフローチャートである。
以下では、本発明に係る動画像復号装置2について、図6~図9を参照して説明する。図6は、本発明に係る動画像復号装置2の構成を示すブロック図である。
続いて、動きベクトル復元部24の構成について、図7を参照して説明する。図7は、動きベクトル復元部24の構成を示すブロック図である。
以下では、動きベクトル復元部24の動作の例について、図8を参照して説明する。図8は、動きベクトル復元部24の動作を示すフローチャートである。
続いて、本発明に係る符号化データ#2のデータ構造について、図9を参照して説明する。
本発明は上記の実施形態に限定されるものではない。以下では、本発明の第1の変形例に係る動画像符号化装置について図10を参照して説明する。
以下では、本発明の変形例に係る動画像符号化装置について説明する。本発明の変形例に係る動画像符号化装置は、動画像符号化装置1における予測ベクトル候補選択部314に代えて、予測ベクトル候補選択部315’を備えている。本変形例に係る動画像符号化装置のその他の構成は、第1の実施形態における動画像符号化装置1と同様である。
以下では、本発明の第2の変形例に係る動画像符号化装置について説明する。本発明の第2の変形例に係る動画像符号化装置は、動画像符号化装置1における予測ベクトル候補選択部314に代えて、予測ベクトル候補選択部316を備えている。本変形例に係る動画像符号化装置のその他の構成は、第1の実施形態における動画像符号化装置1と同様である。
以下では、本発明の第3の変形例に係る動画像符号化装置について説明する。本発明の第3の変形例に係る動画像符号化装置は、動画像符号化装置1における予測ベクトル候補選択部314に代えて、予測ベクトル候補選択部317を備えている。本変形例に係る動画像符号化装置のその他の構成は、第1の実施形態における動画像符号化装置1と同様である。
以下では、本発明の第3’の変形例に係る動画像符号化装置について説明する。本発明の第3’の変形例に係る動画像符号化装置は、動画像符号化装置1における予測ベクトル候補選択部314に代えて、予測ベクトル候補選択部317’を備えている。本変形例に係る動画像符号化装置のその他の構成は、第1の実施形態における動画像符号化装置1と同様である。以下では、本発明の第3’の変形例に係る動画像符号化装置について説明する。
また、本変形例に係る動画像復号装置も、予測ベクトル候補選択部314に換えて予測ベクトル候補選択部317’を備えることを除き、第1の実施形態における動画像復号装置2と同様である。
以下では、本発明の第3’’の変形例に係る動画像符号化装置について説明する。本発明の第3’’の変形例に係る動画像符号化装置は、動画像符号化装置1における予測ベクトル候補選択部314に代えて、予測ベクトル候補選択部317’’を備えている。本変形例に係る動画像符号化装置のその他の構成は、第1の実施形態における動画像符号化装置1と同様である。以下では、本発明の第3’’の変形例に係る動画像符号化装置について説明する。
以下では、本発明の第4の変形例に係る動画像符号化装置について説明する。本発明の第4の変形例に係る動画像符号化装置は、動画像符号化装置1における予測ベクトル候補選択部314に代えて、予測ベクトル候補選択部318を備えている。本変形例に係る動画像符号化装置のその他の構成は、第1の実施形態における動画像符号化装置1と同様である。
上記、実施形態及び変形例では、予測ベクトル候補選択部314(また315、316、317、318など)は、第2の空間的予測ベクトル候補#312、または、時間的予測ベクトル候補#313のうち、何れか一つを選択するとしているが、両者を選択しても構わない。より、具体的には、予測ベクトル候補選択部314において、対象パーティションの時間的予測ベクトル候補#313の予測精度が高いと判定される場合には、第2の空間的予測ベクトル候補#312と時間的予測ベクトル候補#313の両者を選択し、対象パーティションの時間的予測ベクトル候補#313の予測精度が低いと判定される場合には、第2の空間的予測ベクトル候補#312を選択する。
なお、上記、実施形態及び変形例では、第1の空間的予測ベクトル#311と第2次予測ベクトル候補#314を、符号化装置では予測ベクトル選択部32、復号装置では予測ベクトル決定部35において選択する構成としているが、このような構成とせず、第2次予測ベクトル候補#314(これは付記事項1に示したように複数もありうる)をそのまま、予測ベクトル選択部32及び予測ベクトル決定部35で用いる予測ベクトル候補とすることも可能である。
なお、上記、実施形態及び変形例では、スキップブロックでは第2次予測ベクトル候補#314として、常に、第1の空間的予測ベクトル候補#311を用いる構成を説明したがスキップブロックに対してもスキップブロック以外のブロックと同様、時間的予測ベクトル候補#313の予測精度の判定により、第2の空間的予測ベクトル候補#312と時間的予測ベクトル候補#313から選択する構成としても良い。また、付記事項2に示すように第2の空間的予測ベクトル候補#312と時間的予測ベクトル候補#313の両者を選択する構成も構わない。
時間的予測ベクトル候補#313の予測精度の判定の当たりやすさは、シーケンスによって異なることも考えられるため、符号化データ中にどの判定を用いたかを示す情報を符号化する構成とすることも好適である。すなわち、パーティションサイズ、動きベクトル長、フレーム間隔、フレームタイプのどの判定を用いたかを示す情報を符号化する。なお、変形例3’’、変形例4に一例として示したように、複数の判定を組み合わせることも適当である。この場合も組合せを示す情報を符号化する。復号装置では、判定を示す情報を復号し、予測動きベクトル候補の算出、及び、フラグ#19bの復号に利用する。
上記動画像符号化装置は、対象パーティションの左辺に隣接する左パーティション、対象パーティションの上辺に隣接する上パーティション、および、当該上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルのメジアンを、上記対象パーティションに割り付ける予測ベクトルの候補となる第1の予測ベクトル候補に設定する第1の算出手段と、上記左パーティションに割り付けられた動きベクトルを対象パーティションに割り付ける予測ベクトルの候補となる第2の予測ベクトル候補に設定する第2の算出手段と、符号化済フレームにおいて、上記対象パーティションと同じ位置を占めるコロケートパーティションに割り付けられた動きベクトルを上記対象パーティションに割り付ける予測ベクトルの候補となる第3の予測ベクトル候補に設定する第3の算出手段と、上記第3の予測ベクトル候補の予測精度に応じて、上記第2の予測ベクトル候補、または、上記第3の予測ベクトル候補のうち、何れか1つの予測ベクトル候補を選択する第1の選択手段と、符号化コストに応じて、上記第1の予測ベクトル候補、または、上記第1の選択手段によって選択された予測ベクトル候補のうち、何れか1つの予測ベクトル候補を選択し、上記対象パーティションに割り付ける予測ベクトルに設定する第2の選択手段とを備えている、ことが好ましい。
上記対象パーティションのサイズが小さい場合であっても、上記対象パーティションのサイズが大きい場合であっても、符号化効率のより高い予測ベクトル候補を選択することができ、また、上記対象パーティションに割り付けられた動きベクトルの長さが短い場合であっても、長い場合であっても、符号化効率のより高い予測ベクトル候補を選択することができるという更なる効果を奏する。
11 変換・量子化部
12 可変長符号化部
13 逆量子化・逆変換部
14 バッファメモリ
15 イントラ予測画像生成部
16 予測画像生成部
17 動きベクトル推定部
18 予測方式制御部
19 動きベクトル冗長性削減部
21 加算器
22 減算器
31 予測ベクトル候補生成部
311 第1空間的予測ベクトル算出部(第1の算出手段)
312 第2空間的予測ベクトル算出部(第2の算出手段)
313 時間的予測ベクトル算出部(第3の算出手段)
314 予測ベクトル候補選択部(第1の選択手段)
32 予測ベクトル選択部(第2の選択手段)
33 PMVフラグ生成部
34 減算器
2 動画像復号装置
24 動きベクトル復元部
Claims (15)
- 動画像を構成するフレームを分割して得られた複数のパーティションの各々に割り付けられた動きベクトルと予測ベクトルとの差分ベクトルを上記動画像と共に符号化する動画像符号化装置において、
上記予測ベクトルは、
対象パーティションの左辺に隣接する左パーティション、対象パーティションの上辺に隣接する上パーティション、および、当該上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルのメジアンである第1の予測ベクトル候補、または、
上記左パーティションに割り付けられた動きベクトルである第2の予測ベクトル候補、若しくは、復号済フレームにおいて上記対象パーティションと同じ位置を占めるコロケートパーティションに割り付けられた動きベクトルである第3の予測ベクトル候補のうち、上記第3の予測ベクトル候補の予測精度に応じて選択された何れか一つの予測ベクトル候補、
のうち、符号化コストに応じて選択された何れか一つの予測ベクトルである、
ことを特徴とする動画像符号化装置。 - 請求項1に記載の動画像符号化装置であって、
対象パーティションの左辺に隣接する左パーティション、対象パーティションの上辺に隣接する上パーティション、および、当該上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルのメジアンを、上記対象パーティションに割り付ける予測ベクトルの候補となる第1の予測ベクトル候補に設定する第1の算出手段と、
上記左パーティションに割り付けられた動きベクトルを対象パーティションに割り付ける予測ベクトルの候補となる第2の予測ベクトル候補に設定する第2の算出手段と、
符号化済フレームにおいて、上記対象パーティションと同じ位置を占めるコロケートパーティションに割り付けられた動きベクトルを上記対象パーティションに割り付ける予測ベクトルの候補となる第3の予測ベクトル候補に設定する第3の算出手段と、
上記第3の予測ベクトル候補の予測精度に応じて、上記第2の予測ベクトル候補、または、上記第3の予測ベクトル候補のうち、何れか1つの予測ベクトル候補を選択する第1の選択手段と、
符号化コストに応じて、上記第1の予測ベクトル候補、または、上記第1の選択手段によって選択された予測ベクトル候補のうち、何れか1つの予測ベクトル候補を選択し、上記対象パーティションに割り付ける予測ベクトルに設定する第2の選択手段と、を備えている、
ことを特徴とする動画像符号化装置。 - 請求項2に記載の動画像符号化装置であって、
上記第2の選択手段は、上記第1の予測ベクトル候補、または、上記第1の選択手段によって選択された予測ベクトル候補のうち、符号化コストがより小さくなる予測ベクトル候補を選択し、上記対象パーティションに割り付ける予測ベクトルに設定する、
ことを特徴とする動画像符号化装置。 - 請求項2または3に記載の動画像符号化装置であって、
上記第1の選択手段は、上記対象パーティションのサイズが予め定められた基準サイズ以上であるときには、上記第3の予測ベクトル候補を選択し、上記対象パーティションのサイズが上記基準サイズより小さいときには、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像符号化装置。 - 請求項2または3に記載の動画像符号化装置であって、
上記第1の選択手段は、上記対象パーティションに割り付けられた動きベクトルの長さが予め定められた長さ以下であるときには、上記第3の予測ベクトル候補を選択し、上記対象パーティションに割り付けられた動きベクトルの長さが上記予め定められた長さより長いときには、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像符号化装置。 - 請求項2または3に記載の動画像符号化装置であって、
上記第1の選択手段は、上記対象パーティションが含まれるフレームと、上記コロケートパーティションが含まれるフレームとのフレーム間隔が、予め定められた基準フレーム間隔以下であるときには、上記第3の予測ベクトル候補を選択し、上記対象パーティションが含まれるフレームと、上記コロケートパーティションが含まれるフレームとのフレーム間隔が、上記基準フレーム間隔より大きいときには、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像符号化装置。 - 請求項2または3に記載の動画像符号化装置であって、
上記第1の選択手段は、上記対象パーティションが含まれるフレーム、または、上記コロケートパーティションが含まれるフレームのうち、少なくとも一方が、階層Bピクチャによって構成されるピクチャ群構造に属するときには、上記第2の予測ベクトル候補を選択し、上記対象パーティションが含まれるフレーム、および、上記コロケートパーティションが含まれるフレームの双方が、階層Bピクチャによって構成されるピクチャ群構造に属さないときには、上記第3の予測ベクトル候補を選択する、
ことを特徴とする動画像符号化装置。 - 請求項2または3に記載の動画像符号化装置であって、
上記第1の選択手段は、上記対象パーティションのサイズが、予め定められた基準サイズ以上である場合であって、上記対象パーティションに割り付けられた動きベクトルの長さが予め定められた長さ以下である場合には、上記第3の予測ベクトル候補を選択し、それ以外の場合には、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像符号化装置。 - 動画像を構成するフレームを分割して得られた複数のパーティションの各々に割り付けられた動きベクトルと予測ベクトルとの差分ベクトルを上記動画像と共に符号化して得られた符号化データを復号する動画像復号装置であって、
対象パーティションの左辺に隣接する左パーティション、対象パーティションの上辺に隣接する上パーティション、および、当該上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルのメジアンを、上記対象パーティションに割り付ける予測ベクトルの候補となる第1の予測ベクトル候補に設定する第1の算出手段と、
上記左パーティションに割り付けられた動きベクトルを対象パーティションに割り付ける予測ベクトルの候補となる第2の予測ベクトル候補に設定する第2の算出手段と、
復号済フレームにおいて、上記対象パーティションと同じ位置を占めるコロケートパーティションに割り付けられた動きベクトルを上記対象パーティションに割り付ける予測ベクトルの候補となる第3の予測ベクトル候補に設定する第3の算出手段と、
上記第3の予測ベクトル候補の予測精度に応じて、上記第2の予測ベクトル候補、または、上記第3の予測ベクトル候補のうち、何れか1つの予測ベクトル候補を選択する第1の選択手段と、
上記符号化データに含まれるフラグを参照して、上記第1の予測ベクトル候補、または、上記第1の選択手段によって選択された予測ベクトル候補のうち、何れか1つの予測ベクトル候補を選択し、上記対象パーティションに割り付ける予測ベクトルに設定する第2の選択手段と、を備えている、
ことを特徴とする動画像復号装置。 - 請求項9に記載の動画像復号装置であって、
上記第1の選択手段は、上記対象パーティションのサイズが予め定められた基準サイズ以上であるときには、上記第3の予測ベクトル候補を選択し、上記対象パーティションのサイズが上記基準サイズより小さいときには、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像復号装置。 - 請求項9に記載の動画像復号装置であって、
上記第1の選択手段は、上記対象パーティションに割り付けられた動きベクトルの長さが予め定められた長さ以下であるときには、上記第3の予測ベクトル候補を選択し、上記対象パーティションに割り付けられた動きベクトルの長さが上記予め定められた長さより長いときには、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像復号装置。 - 請求項9に記載の動画像復号装置であって、
上記第1の選択手段は、上記対象パーティションが含まれるフレームと、上記コロケートパーティションが含まれるフレームとのフレーム間隔が、予め定められた基準フレーム間隔以下であるときには、上記第3の予測ベクトル候補を選択し、上記対象パーティションが含まれるフレームと、上記コロケートパーティションが含まれるフレームとのフレーム間隔が、上記基準フレーム間隔より大きいときには、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像復号装置。 - 請求項9に記載の動画像復号装置であって、
上記第1の選択手段は、上記符号化データに、符号化された画像が階層Bピクチャによって構成されるピクチャ群構造に属することを示すフラグが含まれている場合には、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像復号装置。 - 請求項9に記載の動画像復号装置であって、
上記第1の選択手段は、上記対象パーティションのサイズが、予め定められた基準サイズ以上である場合であって、上記対象パーティションに割り付けられた動きベクトルの長さが予め定められた長さ以下である場合には、上記第3の予測ベクトル候補を選択し、それ以外の場合には、上記第2の予測ベクトル候補を選択する、
ことを特徴とする動画像復号装置。 - 動画像を構成するフレームを分割して得られた複数のパーティションの各々に割り付けられた動きベクトルと予測ベクトルとの差分ベクトルを上記動画像と共に符号化して得られた符号化データのデータ構造において、
上記予測ベクトルは、
対象パーティションの左辺に隣接する左パーティション、対象パーティションの上辺に隣接する上パーティション、および、当該上パーティションの右辺に隣接する右上パーティションの各々に割り付けられた動きベクトルのメジアンである第1の予測ベクトル候補、または、
上記左パーティションに割り付けられた動きベクトルである第2の予測ベクトル候補、若しくは、復号済フレームにおいて上記対象パーティションと同じ位置を占めるコロケートパーティションに割り付けられた動きベクトルである第3の予測ベクトル候補のうち、上記第3の予測ベクトル候補の予測精度に応じて選択された何れか一つの予測ベクトル候補、
のうち、符号化コストに応じて選択された何れか一つの予測ベクトルである、
ことを特徴とする符号化データのデータ構造。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/502,684 US20120213288A1 (en) | 2009-10-20 | 2010-09-17 | Video encoding device, video decoding device, and data structure |
EP10824754A EP2493195A1 (en) | 2009-10-20 | 2010-09-17 | Video encoding device, video decoding device, and data structure |
JP2011537183A JPWO2011048903A1 (ja) | 2009-10-20 | 2010-09-17 | 動画像符号化装置、動画像復号装置、および、データ構造 |
CN201080046998.9A CN102577392A (zh) | 2009-10-20 | 2010-09-17 | 运动图像编码装置、运动图像解码装置以及数据结构 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2009-241827 | 2009-10-20 | ||
JP2009241827 | 2009-10-20 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2011048903A1 true WO2011048903A1 (ja) | 2011-04-28 |
Family
ID=43900143
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2010/066245 WO2011048903A1 (ja) | 2009-10-20 | 2010-09-17 | 動画像符号化装置、動画像復号装置、および、データ構造 |
Country Status (5)
Country | Link |
---|---|
US (1) | US20120213288A1 (ja) |
EP (1) | EP2493195A1 (ja) |
JP (1) | JPWO2011048903A1 (ja) |
CN (1) | CN102577392A (ja) |
WO (1) | WO2011048903A1 (ja) |
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012090397A1 (ja) * | 2010-12-28 | 2012-07-05 | 株式会社Jvcケンウッド | 動画像符号化装置、動画像符号化方法及び動画像符号化プログラム、並びに動画像復号装置、動画像復号方法及び動画像復号プログラム |
WO2012173022A1 (ja) * | 2011-06-14 | 2012-12-20 | ソニー株式会社 | 画像処理装置および方法 |
JP2013012961A (ja) * | 2011-06-29 | 2013-01-17 | Jvc Kenwood Corp | 画像復号装置、画像復号方法、および画像復号プログラム |
WO2013035452A1 (ja) * | 2011-09-05 | 2013-03-14 | シャープ株式会社 | 画像符号化方法、画像復号方法、並びにそれらの装置及びプログラム |
JP2013517734A (ja) * | 2010-01-19 | 2013-05-16 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
JP2013517669A (ja) * | 2010-01-14 | 2013-05-16 | サムスン エレクトロニクス カンパニー リミテッド | 動きベクトルを符号化/復号化する方法及び装置 |
JP2013528992A (ja) * | 2010-04-22 | 2013-07-11 | 聯發科技股▲ふん▼有限公司 | 動き予測方法 |
WO2013153778A1 (ja) * | 2012-04-09 | 2013-10-17 | パナソニック株式会社 | 画像符号化方法および画像復号化方法 |
WO2014097937A1 (ja) * | 2012-12-20 | 2014-06-26 | ソニー株式会社 | 画像処理装置および画像処理方法 |
JP2014209762A (ja) * | 2010-12-17 | 2014-11-06 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュートElectronics And Telecommunicationsresearch Institute | 映像復号化方法 |
JP2015228680A (ja) * | 2011-09-28 | 2015-12-17 | 株式会社Jvcケンウッド | 動画像復号装置、動画像復号方法、動画像復号プログラム、受信装置、受信方法及び受信プログラム |
JP2016106486A (ja) * | 2011-09-28 | 2016-06-16 | 株式会社Jvcケンウッド | 動画像符号化装置、動画像符号化方法、動画像符号化プログラム、送信装置、送信方法、及び送信プログラム |
TWI569634B (zh) * | 2011-10-31 | 2017-02-01 | Jvc Kenwood Corp | Dynamic image decoding device, dynamic image decoding method and dynamic image decoding program |
JP2017510115A (ja) * | 2014-01-03 | 2017-04-06 | マイクロソフト テクノロジー ライセンシング,エルエルシー | ビデオ及び画像符号化/デコーディングにおけるブロックベクトル予測 |
CN107347160A (zh) * | 2011-11-18 | 2017-11-14 | 谷歌技术控股有限责任公司 | 针对高效率视频代码化用信号发送并置图片的显式方式 |
JP2018026726A (ja) * | 2016-08-10 | 2018-02-15 | 日本放送協会 | 符号化装置及びプログラム |
JP2018088715A (ja) * | 2018-02-23 | 2018-06-07 | マイクロソフト テクノロジー ライセンシング,エルエルシー | イントラブロックコピー予測におけるブロック反転及びスキップモード |
JP2018093504A (ja) * | 2010-01-19 | 2018-06-14 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
US10368091B2 (en) | 2014-03-04 | 2019-07-30 | Microsoft Technology Licensing, Llc | Block flipping and skip mode in intra block copy prediction |
US10390034B2 (en) | 2014-01-03 | 2019-08-20 | Microsoft Technology Licensing, Llc | Innovations in block vector prediction and estimation of reconstructed sample values within an overlap area |
US10582213B2 (en) | 2013-10-14 | 2020-03-03 | Microsoft Technology Licensing, Llc | Features of intra block copy prediction mode for video and image coding and decoding |
US10785486B2 (en) | 2014-06-19 | 2020-09-22 | Microsoft Technology Licensing, Llc | Unified intra block copy and inter prediction modes |
US10812817B2 (en) | 2014-09-30 | 2020-10-20 | Microsoft Technology Licensing, Llc | Rules for intra-picture prediction modes when wavefront parallel processing is enabled |
US10986349B2 (en) | 2017-12-29 | 2021-04-20 | Microsoft Technology Licensing, Llc | Constraints on locations of reference blocks for intra block copy prediction |
US11109036B2 (en) | 2013-10-14 | 2021-08-31 | Microsoft Technology Licensing, Llc | Encoder-side options for intra block copy prediction mode for video and image coding |
US11284103B2 (en) | 2014-01-17 | 2022-03-22 | Microsoft Technology Licensing, Llc | Intra block copy prediction with asymmetric partitions and encoder-side search patterns, search ranges and approaches to partitioning |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2493755B (en) * | 2011-08-17 | 2016-10-19 | Canon Kk | Method and device for encoding a sequence of images and method and device for decoding a sequence of images |
GB2556489B (en) | 2011-11-08 | 2018-11-21 | Kt Corp | A method of decoding a video signal using a merge mode |
CN111526360A (zh) * | 2016-02-06 | 2020-08-11 | 华为技术有限公司 | 图像编解码方法及装置 |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008283490A (ja) * | 2007-05-10 | 2008-11-20 | Ntt Docomo Inc | 動画像符号化装置、方法及びプログラム、並びに動画像復号化装置、方法及びプログラム |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6765963B2 (en) * | 2001-01-03 | 2004-07-20 | Nokia Corporation | Video decoder architecture and method for using same |
EP1378124B1 (en) * | 2001-02-13 | 2005-06-01 | Koninklijke Philips Electronics N.V. | Motion information coding and decoding method |
CN100493201C (zh) * | 2002-01-24 | 2009-05-27 | 株式会社日立制作所 | 运动图像的解码方法 |
-
2010
- 2010-09-17 WO PCT/JP2010/066245 patent/WO2011048903A1/ja active Application Filing
- 2010-09-17 US US13/502,684 patent/US20120213288A1/en not_active Abandoned
- 2010-09-17 EP EP10824754A patent/EP2493195A1/en not_active Withdrawn
- 2010-09-17 JP JP2011537183A patent/JPWO2011048903A1/ja active Pending
- 2010-09-17 CN CN201080046998.9A patent/CN102577392A/zh active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008283490A (ja) * | 2007-05-10 | 2008-11-20 | Ntt Docomo Inc | 動画像符号化装置、方法及びプログラム、並びに動画像復号化装置、方法及びプログラム |
Non-Patent Citations (1)
Title |
---|
JOEL JUNG ET AL.: "Competition-Based Scheme for Motion Vector Selection and Coding", ITU-TELECOMMUNICATIONS STANDARDIZATION SECTOR STUDY GROUP 16 QUESTION 6 VIDEO CODING EXPERTS GROUP (VCEG) 29TH MEETING VCEG-AC06,ITU, - 17 July 2006 (2006-07-17), XP030003490 * |
Cited By (56)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2013517669A (ja) * | 2010-01-14 | 2013-05-16 | サムスン エレクトロニクス カンパニー リミテッド | 動きベクトルを符号化/復号化する方法及び装置 |
US9131237B2 (en) | 2010-01-14 | 2015-09-08 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector by predicting motion vector according to mode |
US9106924B2 (en) | 2010-01-14 | 2015-08-11 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector by predicting motion vector according to mode |
US8995529B2 (en) | 2010-01-14 | 2015-03-31 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector by predicting motion vector according to mode |
JP2013517734A (ja) * | 2010-01-19 | 2013-05-16 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
JP2016007065A (ja) * | 2010-01-19 | 2016-01-14 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
JP2019146252A (ja) * | 2010-01-19 | 2019-08-29 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
US10057597B2 (en) | 2010-01-19 | 2018-08-21 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector based on reduced motion vector predictor candidates |
JP2018093504A (ja) * | 2010-01-19 | 2018-06-14 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
JP2017108423A (ja) * | 2010-01-19 | 2017-06-15 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
US10484709B2 (en) | 2010-01-19 | 2019-11-19 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector based on reduced motion vector predictor candidates |
US10708616B2 (en) | 2010-01-19 | 2020-07-07 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector based on reduced motion vector predictor candidates |
JP2020182236A (ja) * | 2010-01-19 | 2020-11-05 | サムスン エレクトロニクス カンパニー リミテッド | 縮小された予測動きベクトルの候補に基づいて、動きベクトルを符号化/復号化する方法及び装置 |
US9571853B2 (en) | 2010-01-19 | 2017-02-14 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector based on reduced motion vector predictor candidates |
US9565447B2 (en) | 2010-01-19 | 2017-02-07 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding motion vector based on reduced motion vector predictor candidates |
JP2013528992A (ja) * | 2010-04-22 | 2013-07-11 | 聯發科技股▲ふん▼有限公司 | 動き予測方法 |
US10708614B2 (en) | 2010-12-17 | 2020-07-07 | Electronics And Telecommunications Research Institute | Method and apparatus for inter prediction using motion vector candidate based on temporal motion prediction |
JP2018038050A (ja) * | 2010-12-17 | 2018-03-08 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュートElectronics And Telecommunications Research Institute | 映像復号化方法 |
JP2016158296A (ja) * | 2010-12-17 | 2016-09-01 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュートElectronics And Telecommunications Research Institute | 映像復号化方法 |
US10397599B2 (en) | 2010-12-17 | 2019-08-27 | Electronics And Telecommunications Research Institute | Method and apparatus for inter prediction using motion vector candidate based on temporal motion prediction |
JP2022040120A (ja) * | 2010-12-17 | 2022-03-10 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュート | 映像復号化方法 |
JP2020174362A (ja) * | 2010-12-17 | 2020-10-22 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュートElectronics And Telecommunications Research Institute | 映像復号化方法 |
JP6992128B2 (ja) | 2010-12-17 | 2022-01-13 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュート | 映像復号化方法 |
JP2014209762A (ja) * | 2010-12-17 | 2014-11-06 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュートElectronics And Telecommunicationsresearch Institute | 映像復号化方法 |
JP2019110595A (ja) * | 2010-12-17 | 2019-07-04 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュートElectronics And Telecommunications Research Institute | 映像復号化方法 |
US11206424B2 (en) | 2010-12-17 | 2021-12-21 | Electronics And Telecommunications Research Institute | Method and apparatus for inter prediction using motion vector candidate based on temporal motion prediction |
JP7448694B2 (ja) | 2010-12-17 | 2024-03-12 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュート | 映像復号化方法 |
JP7210687B2 (ja) | 2010-12-17 | 2023-01-23 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュート | 映像復号化方法 |
US11743486B2 (en) | 2010-12-17 | 2023-08-29 | Electronics And Telecommunications Research Institute | Method and apparatus for inter prediction using motion vector candidate based on temporal motion prediction |
WO2012090397A1 (ja) * | 2010-12-28 | 2012-07-05 | 株式会社Jvcケンウッド | 動画像符号化装置、動画像符号化方法及び動画像符号化プログラム、並びに動画像復号装置、動画像復号方法及び動画像復号プログラム |
CN103597836A (zh) * | 2011-06-14 | 2014-02-19 | 索尼公司 | 图像处理设备和方法 |
WO2012173022A1 (ja) * | 2011-06-14 | 2012-12-20 | ソニー株式会社 | 画像処理装置および方法 |
JP2013012961A (ja) * | 2011-06-29 | 2013-01-17 | Jvc Kenwood Corp | 画像復号装置、画像復号方法、および画像復号プログラム |
WO2013035452A1 (ja) * | 2011-09-05 | 2013-03-14 | シャープ株式会社 | 画像符号化方法、画像復号方法、並びにそれらの装置及びプログラム |
JP2015228680A (ja) * | 2011-09-28 | 2015-12-17 | 株式会社Jvcケンウッド | 動画像復号装置、動画像復号方法、動画像復号プログラム、受信装置、受信方法及び受信プログラム |
JP2016106486A (ja) * | 2011-09-28 | 2016-06-16 | 株式会社Jvcケンウッド | 動画像符号化装置、動画像符号化方法、動画像符号化プログラム、送信装置、送信方法、及び送信プログラム |
TWI597973B (zh) * | 2011-10-31 | 2017-09-01 | Jvc Kenwood Corp | Video encoding device, video encoding method and recording medium |
TWI569634B (zh) * | 2011-10-31 | 2017-02-01 | Jvc Kenwood Corp | Dynamic image decoding device, dynamic image decoding method and dynamic image decoding program |
CN107347160A (zh) * | 2011-11-18 | 2017-11-14 | 谷歌技术控股有限责任公司 | 针对高效率视频代码化用信号发送并置图片的显式方式 |
CN107347160B (zh) * | 2011-11-18 | 2020-04-28 | 谷歌技术控股有限责任公司 | 针对高效率视频代码化用信号发送并置图片的显式方式 |
US10075708B2 (en) | 2012-04-09 | 2018-09-11 | Sun Patent Trust | Image encoding method and image decoding method |
JPWO2013153778A1 (ja) * | 2012-04-09 | 2015-12-17 | パナソニック インテレクチュアル プロパティ コーポレーション オブアメリカPanasonic Intellectual Property Corporation of America | 画像符号化方法および画像復号化方法 |
WO2013153778A1 (ja) * | 2012-04-09 | 2013-10-17 | パナソニック株式会社 | 画像符号化方法および画像復号化方法 |
WO2014097937A1 (ja) * | 2012-12-20 | 2014-06-26 | ソニー株式会社 | 画像処理装置および画像処理方法 |
US11109036B2 (en) | 2013-10-14 | 2021-08-31 | Microsoft Technology Licensing, Llc | Encoder-side options for intra block copy prediction mode for video and image coding |
US10582213B2 (en) | 2013-10-14 | 2020-03-03 | Microsoft Technology Licensing, Llc | Features of intra block copy prediction mode for video and image coding and decoding |
US10390034B2 (en) | 2014-01-03 | 2019-08-20 | Microsoft Technology Licensing, Llc | Innovations in block vector prediction and estimation of reconstructed sample values within an overlap area |
JP2017510115A (ja) * | 2014-01-03 | 2017-04-06 | マイクロソフト テクノロジー ライセンシング,エルエルシー | ビデオ及び画像符号化/デコーディングにおけるブロックベクトル予測 |
US10469863B2 (en) | 2014-01-03 | 2019-11-05 | Microsoft Technology Licensing, Llc | Block vector prediction in video and image coding/decoding |
US11284103B2 (en) | 2014-01-17 | 2022-03-22 | Microsoft Technology Licensing, Llc | Intra block copy prediction with asymmetric partitions and encoder-side search patterns, search ranges and approaches to partitioning |
US10368091B2 (en) | 2014-03-04 | 2019-07-30 | Microsoft Technology Licensing, Llc | Block flipping and skip mode in intra block copy prediction |
US10785486B2 (en) | 2014-06-19 | 2020-09-22 | Microsoft Technology Licensing, Llc | Unified intra block copy and inter prediction modes |
US10812817B2 (en) | 2014-09-30 | 2020-10-20 | Microsoft Technology Licensing, Llc | Rules for intra-picture prediction modes when wavefront parallel processing is enabled |
JP2018026726A (ja) * | 2016-08-10 | 2018-02-15 | 日本放送協会 | 符号化装置及びプログラム |
US10986349B2 (en) | 2017-12-29 | 2021-04-20 | Microsoft Technology Licensing, Llc | Constraints on locations of reference blocks for intra block copy prediction |
JP2018088715A (ja) * | 2018-02-23 | 2018-06-07 | マイクロソフト テクノロジー ライセンシング,エルエルシー | イントラブロックコピー予測におけるブロック反転及びスキップモード |
Also Published As
Publication number | Publication date |
---|---|
US20120213288A1 (en) | 2012-08-23 |
EP2493195A1 (en) | 2012-08-29 |
CN102577392A (zh) | 2012-07-11 |
JPWO2011048903A1 (ja) | 2013-03-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2011048903A1 (ja) | 動画像符号化装置、動画像復号装置、および、データ構造 | |
US8873633B2 (en) | Method and apparatus for video encoding and decoding | |
JP6558477B2 (ja) | 映像符号化装置及び映像復号装置 | |
KR101383540B1 (ko) | 복수의 움직임 벡터 프리딕터들을 사용하여 움직임 벡터를추정하는 방법, 장치, 인코더, 디코더 및 복호화 방법 | |
JP6100240B2 (ja) | 多視点ビデオの動きベクトル符号化方法及び装置、その復号化方法及び装置 | |
JP5061179B2 (ja) | 照明変化補償動き予測符号化および復号化方法とその装置 | |
US20120207221A1 (en) | Video coding device and video decoding device | |
KR20080088040A (ko) | 영상의 부호화, 복호화 방법 및 장치 | |
TW202017379A (zh) | 動態影像編碼裝置及動態影像編碼方法 | |
JP2013225892A (ja) | 直接モード符号化及び復号化装置 | |
JP2007329693A (ja) | 画像符号化装置、及び画像符号化方法 | |
JPWO2010070818A1 (ja) | 動画像符号化装置、動画像符号化方法、動画像復号化装置および動画像復号化方法 | |
KR101238012B1 (ko) | 예측 움직임 벡터 부호화 방법 및 이를 수행하는 부호화 장치 | |
JP2009005217A (ja) | 画像符号化方法,画像復号方法,画像符号化装置,画像復号装置,画像符号化プログラム,画像復号プログラムおよびコンピュータ読み取り可能な記録媒体 | |
JP4865767B2 (ja) | スケーラブル動画像符号化方法、スケーラブル動画像符号化装置、スケーラブル動画像符号化プログラムおよびそのプログラムを記録したコンピュータ読み取り可能な記録媒体 | |
JP4243205B2 (ja) | 動画像符号化装置、その方法及びそのプログラム | |
KR101315295B1 (ko) | 다시점 영상의 부호화, 복호화 방법 및 장치 | |
JP2002232892A (ja) | 画像符号化装置 | |
JP2018026726A (ja) | 符号化装置及びプログラム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201080046998.9 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 10824754 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2011537183 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13502684 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010824754 Country of ref document: EP |