JP6917718B2 - Predictors, encoders, decoders, and programs - Google Patents

Description

The present invention relates to a prediction device, a coding device, a decoding device, and a program, and more particularly to an in-screen prediction method for coding and decoding an image.

MPEG-2 and H.M. 264 / AVC, H. In 265 / HEVC, each frame of a moving image is divided into square (rectangular) blocks, and each block is used as a basic processing unit to generate a predicted image. Then, the difference between the original image and the predicted image is encoded.
There are two types of generation methods for generating a prediction image: intra (in-screen) prediction and inter (inter-screen) prediction. In the intra-prediction, extrapolation prediction is performed using the peripheral pixels of the prediction target block as reference pixels. In the inter-prediction, interpolation prediction or extrapolation prediction is performed from the frames before and after the prediction target frame.

In addition, H. In 265 / HEVC, there are three types of intra-prediction methods: DC (Direct Current) prediction, Planar prediction, and direction prediction (for example, Non-Patent Document 1). Of these, DC prediction is a method of calculating the average value of reference pixels and using the average value as the prediction value to generate a prediction image of the prediction target block. For example, the prediction target block _{(P (1,1) ~P (N} , N)) shown in FIG. 11 of the case, the prediction image of the DC prediction all the pixel values are predicted value _{P DC} N × N pixels of the image It becomes. Here, the predicted value P _DC is an average value of reference pixels (R _{(0,0) to} R _{(0, N)} , R _{(1,0) to} R _{(N, 0)} ), and is expressed by the following equation ( Calculate from 1).

The size of the prediction target block has been expanded with each standardization of new prediction coding schemes. For example, the size of the block to be predicted is a maximum of 16 × 16 pixels in MPEG-2, and H.I. In 265 / HEVC, the maximum is 64 × 64 pixels.

Recognition ITU-T H. 265, (04/2015), "High efficiency video coding", International Telecommunication Union

By the way, in the conventional DC prediction, the predicted value of each pixel in the prediction target block is uniformly set as the average value of the reference pixels. However, the pixel value of each pixel in the prediction target block is not always close to the average value of the reference pixels. Further, when the prediction target block is large, the pixel values of each pixel in the prediction target block are difficult to be uniform, and the pixel values of some pixels may deviate significantly from the prediction value. Therefore, it may not be possible to improve the coding efficiency of DC prediction.

One of the objects of the present invention is to provide a prediction device, a coding device, a decoding device, and a program capable of improving the coding efficiency of DC prediction.

Another object of the present invention is to provide a prediction device, a coding device, a decoding device, and a program capable of exerting the effects described in the embodiments described later.

In order to solve the above-mentioned problems, one aspect of the present invention is a specific unit that specifies a plurality of reference pixel sets for one image block, and a set of a plurality of reference pixels specified by the specific unit. A prediction device including a generation unit that generates a prediction block candidate by DC prediction and a selection unit that selects a prediction block for the image block from a plurality of prediction block candidates generated by the generation unit. Is.

Further, one aspect of the present invention is the above-mentioned prediction device, and the specific unit specifies a set composed of adjacent reference pixels among a plurality of reference pixels.
Further, one aspect of the present invention is the above-mentioned prediction device, and the specific part specifies a predetermined number of sets.
Further, one aspect of the present invention is the above-mentioned prediction device, in which the specific unit specifies a set of reference pixels by clustering.
Further, one aspect of the present invention is the above-mentioned prediction device, and the selection unit selects a prediction block for each of the divided blocks obtained by dividing the image block into a plurality of the prediction block candidates.

Further, one aspect of the present invention is a coding device including a coding unit that encodes flag information indicating a prediction block selected by the prediction device.
Another embodiment of the present invention, the encoding device based on the encoded flag information, a decoding apparatus comprising a decrypt unit for decrypting the image block.
Further, one aspect of the present invention is a program for operating a computer as the above-mentioned prediction device.

According to one aspect of the present invention, the coding efficiency of DC prediction can be improved.

It is a figure which shows the outline of DC prediction which concerns on 1st Embodiment of this invention. It is a block diagram which shows the structure of the image coding decoding system which concerns on the same embodiment. It is a block diagram which shows the structure of the coding apparatus which concerns on this embodiment. It is a block diagram which shows the structure of the intra prediction part which concerns on this embodiment. It is a block diagram which shows the structure of the decoding apparatus which concerns on this embodiment. It is a flowchart which shows the operation of the intra prediction part which concerns on this embodiment. It is a figure which shows the outline of DC prediction which concerns on 2nd Embodiment of this invention. It is a block diagram which shows the structure of the intra prediction part which concerns on this embodiment. It is a flowchart which shows the operation of the intra prediction part which concerns on this embodiment. It is a figure which shows an example of the processing result by the image coding decoding system which concerns on the same embodiment. It is a figure which shows the outline of DC prediction which concerns on the prior art.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
[First Embodiment]
The first embodiment of the present invention will be described. First, an outline of the image coding / decoding system 1 according to the present embodiment will be described. The image coding / decoding system 1 is an information processing system that performs intra-prediction such as DC prediction and inter-prediction to encode and decode an image. The image coding / decoding system 1 may process a still image or a moving image (video) as a processing target. In the following, the image is represented by a two-dimensional Cartesian coordinate system having a horizontal direction and a vertical direction. Specifically, the coordinates of the image in the two-dimensional Cartesian coordinate system are represented by two components such as (x, y), the first component (x component) is the horizontal coordinate, and the second component (y component) is the vertical coordinate. Define. Further, the pixel value at the coordinates (x, y) is expressed in the format of D (x, y).

Here, the outline of the DC prediction according to the present embodiment will be described.
FIG. 1 is a diagram showing an outline of DC prediction according to the present embodiment.
In the example shown in FIG. 1, the image block (hereinafter referred to as the prediction target block) P for which DC prediction is performed has the size of (N-1) × (N-1) pixels (N is positive). integer). The shape of the prediction target block P is a rectangle such as a square or a rectangle. Pixels (hereinafter, referred to as reference pixels) R referred to when performing DC prediction are present on the left side and the upper side of the prediction target block P with respect to the prediction target block P. In the example shown in FIG. 1, the reference pixel R (0,0) is a pixel located at a virtual origin. In DC prediction according to the prior art in Non-Patent Document 1, the pixel values of the coordinates (0,0) to (0,N) adjacent to the prediction target block P and the coordinates (1,0) to (N,0). One predicted value was generated by referring to R _{(0,0) to} R _{(0, N)} and R _{(1,0) to} R _{(N, 0).}

On the other hand, in the DC prediction according to the present embodiment, the plurality of reference pixels R are classified into a plurality of groups. Each group contains one or more pixels. In the example shown in FIG. 1, a plurality of reference pixels R are divided into _{four groups, DC A1} , DC _A2 , DC _L1 , and DC _{L2, in the first to fourth groups.} Specifically, the first group DC _A1 includes pixels at coordinates (1,0) to (K, 0). The second group DC _A2 includes pixels of coordinates (K + 1,0) to (N,0). The third group DC _L1 includes pixels having coordinates (0, 1) to (0, L). The fourth group DC _L2 includes pixels having coordinates (0, L + 1) to (0, N). In this way, the image coding / decoding system 1 specifies a group of reference pixels R.

Next, the image coding / decoding system 1 makes a DC prediction with reference to the reference pixel R included in each group. That is, the predicted value of the DC prediction can be obtained for each group. The predicted value for each group is a candidate for the predicted value to be encoded, and will be referred to as a predicted candidate value below. The image coding / decoding system 1 selects any one of the plurality of prediction candidate values as the prediction value. Then, the image coding / decoding system 1 encodes and decodes the flag information indicating the selection result (hereinafter, referred to as the selection prediction value).

In this way, the image coding / decoding system 1 acquires a plurality of prediction candidate values by DC prediction for one prediction target block P. Then, for coding and decoding, one selected prediction value out of a plurality of prediction candidate values is used. That is, the image coding / decoding system 1 can use a candidate close to the prediction target block P as the selection prediction value. Therefore, the image coding / decoding system 1 can improve the coding accuracy as compared with the case where only one predicted value by DC prediction is acquired for one prediction target block P.
The above is a description of the outline of the image coding / decoding system 1.

Next, the configuration of the image coding / decoding system 1 will be described.
FIG. 2 is a block diagram showing the configuration of the image coding / decoding system 1.
The image coding / decoding system 1 includes a coding device 10, a decoding device 30, and a transmission storage unit 50. The coding device 10 and the decoding device 30 are, for example, electronic devices including a computer system.
The coding device 10 encodes the image and outputs it to the transmission storage unit 50. That is, the coding device 10 outputs a code.
The decoding device 30 acquires the code output by the coding device 10. The decoding device 30 decodes the code and outputs an image.

The transmission storage unit 50 transmits or stores the code. The transmission storage unit 50 performs transmission line coding and signal modulation at the transmission end as needed. Then, the transmission storage unit 50 transmits a signal through a transmission line of wireless, wired, or a combination thereof. At the receiving end of the transmission storage unit 50, demodulation and transmission line decoding are performed as necessary, and a code is output. The transmission storage unit 50 may perform encryption, add an electronic signature, and perform multiplexing or demultiplexing with other signals. Further, the transmission storage unit 50 may perform distribution, exchange, remodulation, delay, amplification, etc. in the middle path of the transmission line.

Further, when the transmission storage unit 50 stores the code, the transmission storage unit 50 stores the acquired code in the storage medium. The transmission storage unit 50 may output the code at the same time as the code is stored, or may read and output the code once stored. When the acquired code is stored in the storage medium, the transmission storage unit 50 may perform coding, encryption, addition of an electronic signature, or the like with an error correction code or the like. The storage medium may be a solid-state memory, or may be a magnetic or optical recording medium.

Further, the transmission storage unit 50 may be in a form in which transmission and storage are combined. In this case, the transmission storage unit 50 transmits the code to a remote storage medium, such as a so-called cloud, and stores the code in the remote storage medium. Then, the transmission storage unit 50 transmits the code to the decoding device 30 at the same time as the code is stored, or reads the stored code and outputs it to the decoding device 30.
Here, it is assumed that the transmission storage unit 50 does not cause a signal error or signal deterioration in the process of transmission and storage. That is, the transmission storage unit 50 outputs the code acquired from the coding device 10 to the decoding device 30 as it is.
The above is the description of the configuration of the image coding / decoding system 1.

Next, the configuration of the coding device 10 will be described.
FIG. 3 is a block diagram showing the configuration of the coding device 10.
The coding device 10 includes a screen division unit 101, a subtraction unit 102, a conversion unit 103, a quantization unit 104, an entropy coding unit 105, an inverse quantization unit 106, an inverse conversion unit 107, and an addition unit. It includes 108, a loop filter 109, an intra prediction unit 200, a motion compensation prediction unit 110, and a switching unit 111. These configurations are realized, for example, by the CPU (Central Processing Unit, not shown) included in the coding device 10 executing a program stored in advance in the storage unit (not shown) included in the coding device 10. .. However, these configurations may be realized as hardware such as an ASIC (Application Specific Integrated Circuit).

The screen dividing unit 101 acquires an image signal to be encoded and divides the image signal into blocks of processing units for coding processing. The screen division unit 101 outputs the signal of the divided block to the subtraction unit 102.
The subtracting unit 102 subtracts the signal acquired from the switching unit 111 from the signal acquired from the screen dividing unit 101. The subtraction unit 102 outputs the signal of the subtraction result to the conversion unit 103.

The conversion unit 103 performs orthogonal conversion on the signal acquired from the subtraction unit 102, and outputs the signal resulting from the orthogonal conversion to the quantization unit 104. The orthogonal transform performed by the transforming unit 103 may be either DCT (Discrete Cosine Transform) or DST (Discrete Sine Transform), or may be both DCT and DST. When the conversion unit 103 performs both DCT and DST, the conversion unit 103 selects either the DCT result or the DST result and outputs it to the quantization unit 104. When the conversion unit 103 selects either the DCT result or the DST result, the conversion unit 103 selects, for example, the one having the better coding efficiency. The signal output as a result of the DCT is the DCT coefficient. The signal output as a result of DST is the DST coefficient.

The quantization unit 104 quantizes the signal acquired from the conversion unit 103, and outputs the quantized signal to the entropy coding unit 105 and the inverse quantization unit 106.
The entropy coding unit 105 performs entropy coding on the signal acquired from the quantization unit 104 and outputs the coded signal. Further, the entropy encoding unit 105 acquires flag information indicating the selected prediction value from the intra prediction unit 200. Then, entropy coding is performed on the flag information, and the coded signal is output.

The inverse quantization unit 106 performs inverse quantization on the signal acquired from the quantization unit 104, and outputs the inverse quantization signal to the inverse conversion unit 107.
The inverse conversion unit 107 performs inverse orthogonal transformation on the signal acquired from the inverse quantization unit 106, and outputs the signal resulting from the inverse orthogonal transformation to the addition unit 108. The inverse orthogonal transformation performed by the inverse conversion unit 107 is an inverse transformation of the orthogonal transformation performed by the conversion unit 103.
The addition unit 108 adds the signal acquired from the inverse conversion unit 107 and the signal acquired from the switching unit 111, and outputs the signal of the addition result to the loop filter 109.
The loop filter 109 performs filtering such as smoothing on the signal acquired from the addition unit 108, and outputs the filtering result signal to the intra prediction unit 200 and the motion compensation prediction unit 110.

The intra prediction unit 200 performs intra prediction with reference to the signal acquired from the loop filter 109, and outputs the signal of the prediction image to the switching unit 111. Here, the configuration of the intra prediction unit 200 will be described.
FIG. 4 is a block diagram showing the configuration of the intra prediction unit 200.
The intra prediction unit 200 includes a Planar prediction unit 201, a direction prediction unit 202, and a DC prediction unit 203.
Planar prediction unit 201 makes Planar prediction.
The direction prediction unit 202 predicts the direction.
The DC prediction unit 203 makes a DC prediction. The DC prediction unit 203 includes a specific unit 204, a generation unit 205, and a selection unit 206.

The identification unit 204 specifies a plurality of reference pixel groups for each prediction target block P. That is, the specifying unit 204 specifies the reference pixels that make up each group. The method of specifying the reference pixels constituting each group is predetermined, and the same group of reference pixels is used for the same prediction target block regardless of whether the coding device 10 or the decoding device 30 is used. Is identified. The same applies to the method of assigning the identification information of the group. For example, the specific unit 204 specifies the pixel having the smallest coordinate value in the horizontal direction and the vertical direction from each group as a representative value, and compares the representative values with each other. Then, a small identification number is assigned to the group including the pixels having the small coordinate values in the horizontal direction. When the coordinate values in the horizontal direction are the same, a small identification number is assigned to the group including the pixels having the small coordinate values in the vertical direction. The method of identifying the reference pixels constituting each group and the method of allocating the identification information of the group may be any method.
The specific unit 204 outputs the group information related to the reference pixels constituting each group to the generation unit 205. The group information includes, for example, group identification information and information indicating a range of reference pixels constituting each group.

The generation unit 205 refers to the reference pixels of each group specified by the specific unit 204 and performs DC prediction. Specifically, the average value of the pixel values of the reference pixels constituting each group is calculated. Then, the generation unit 205 outputs the information of the prediction candidate value acquired from the reference pixel of each group to the selection unit 206.

The selection unit 206 acquires information on the prediction candidate value for each group from the generation unit 205. The selection unit 206 selects one selection prediction value from a plurality of prediction candidate values. For example, the selection unit 206 calculates the difference between the image signal of the prediction target block and the image signal of the prediction image based on the prediction candidate value, and sets the prediction candidate value having the smallest difference as the selection prediction value. The image signal of the prediction image composed of the prediction candidate values is a signal having the same size as the prediction target block and the pixel value of each pixel being the prediction candidate value. The selection unit 206 outputs the signal of the prediction image composed of the selection prediction values to the subtraction unit 102 and the addition unit 108. The signal of the prediction image composed of the selection prediction value is a signal having the same size as the prediction target block and the pixel value of each pixel is the selection prediction value. Further, the selection unit 206 outputs flag information indicating the selection prediction value to the entropy coding unit 105. The flag information is, for example, identification information of a group of reference pixels.

The motion compensation prediction unit 110 makes motion compensation prediction with reference to the signal acquired from the loop filter 109, and outputs the signal of the prediction image to the switching unit 111.
The switching unit 111 switches whether to output the signal acquired from the intra prediction unit 200 or the signal acquired from the motion compensation prediction unit 110. The signal output from the switching unit 111 is input to the subtraction unit 102 and the addition unit 108.
The above is the description of the configuration of the coding device 10.

Next, the configuration of the decoding device 30 will be described.
FIG. 5 is a block diagram showing the configuration of the decoding device 30.
The decoding device 30 performs decoding corresponding to the coding of the coding device 10. The decoding device 30 includes an entropy decoding unit 301, an inverse quantization unit 302, an inverse conversion unit 303, an addition unit 304, a screen coupling unit 305, a loop filter 306, an intra prediction unit 200, and a motion prediction unit 307. And a switching unit 308 is provided. These configurations are realized, for example, by the CPU (not shown) included in the decoding device 30 executing a program stored in advance in the storage unit (not shown) included in the decoding device 30. However, these configurations may be realized as hardware such as ASIC.

The entropy decoding unit 301 acquires a coded signal to be decoded, and performs entropy decoding on the acquired coded signal. The entropy decoding unit 301 outputs the signal of the entropy decoding result to the inverse quantization unit 302.
The dequantization unit 302 performs dequantization on the signal acquired from the entropy decoding unit 301, and outputs the dequantized signal to the reverse conversion unit 303. The dequantization unit 302 has the same configuration as the dequantization unit 106 of the coding device 10.

The inverse conversion unit 303 performs inverse orthogonal transformation on the signal acquired from the inverse quantization unit 302, and outputs the signal resulting from the inverse orthogonal transformation to the addition unit 304. The inverse conversion unit 303 has the same configuration as the inverse conversion unit 107 of the coding device 10.
The addition unit 304 adds the signal acquired from the inverse conversion unit 303 and the signal acquired from the switching unit 308, and outputs the signal of the addition result to the screen coupling unit 305 and the loop filter 306.
The screen coupling unit 305 generates a video signal by combining blocks of processing units of decoding processing composed of signals acquired from the addition unit 304. The screen coupling unit 305 outputs the generated video signal as a decoding signal.

The loop filter 306 performs filtering such as smoothing on the signal acquired from the addition unit 304, and outputs the filtering result signal to the intra prediction unit 200 and the motion prediction unit 307. The loop filter 306 has the same configuration as the loop filter 109 of the coding device 10.
The intra prediction unit 200 performs intra prediction with reference to the signal acquired from the loop filter 306, and outputs the signal of the prediction image to the switching unit 308. The intra prediction unit 200 of the decoding device 30 has the same configuration as the intra prediction unit 200 of the coding device 10. However, the intra prediction unit 200 of the decoding device 30 specifies the selected prediction value with reference to the flag information decoded by the entropy decoding unit 301.

The motion prediction unit 307 performs motion prediction with reference to the signal acquired from the loop filter 306, and outputs the signal of the predicted image to the switching unit 308. The motion prediction unit 307 has the same configuration as the motion compensation prediction unit 110.
The switching unit 308 switches whether to output the signal acquired from the intra prediction unit 200 or the signal acquired from the motion prediction unit 307. The switching control of the switching unit 308 is performed based on the addition result of the addition unit 304. The signal output from the switching unit 308 is input to the adding unit 304. The switching unit 308 has the same configuration as the switching unit 111.
The above is the description of the configuration of the decoding device 30.

Next, the flow of the DC prediction process by the DC prediction unit 203 will be described.
FIG. 6 is a flowchart showing the operation of the DC prediction unit 203.
Here, as an example, the processing flow of the DC prediction unit 203 in the coding apparatus 10 will be described.
(Step S1) The DC prediction unit 203 specifies a group of reference pixels. Next, the DC prediction unit 203 proceeds to step S3.

Here, the identification of the group will be described.
For example, the DC prediction unit 203 divides the reference pixel located on the left side of the prediction target block P based on the coordinate values in the vertical direction. Further, for example, the DC prediction unit 203 divides the reference pixel located above the prediction target block based on the coordinate values in the horizontal direction. When the reference pixel is divided into two in the horizontal direction and the vertical direction, four groups are specified as shown in FIG.

The method for specifying the group is not limited to the method described here. The DC prediction unit 203 may divide the reference pixel R only in either the horizontal direction or the vertical direction, for example. As described above, the number of groups may be any number of 2 or more. In addition to grouping the reference pixels R based on the coordinate values, clustering may be performed based on the pixel values, for example. As a clustering method, for example, the k-means method can be applied. When the k-means method is applied, the number of nodes may be the number of reference pixels R, and the number of clusters may be the number of groups. Further, the DC prediction unit 203 does not necessarily have to include all of the reference pixels R in any of the groups. That is, among the reference pixels R referred to in the conventional DC prediction, there may be pixels that are not included in the group. Further, one pixel may be included in a plurality of groups in an overlapping manner.

Further, the group may be composed of only one pixel. For example, among the reference pixels, the lower left pixel (for example, the coordinates (0, N) in FIG. 1), the upper right pixel (for example, the coordinates (N, 0) in FIG. 1), and the upper left pixel (for example,). The coordinates (0,0) in FIG. 1) may be selected to form three groups. Further, the pixel having the maximum pixel value and the pixel having the minimum pixel value may be selected from the reference pixels R to form different groups. Further, a group including all the pixels of the reference pixel R may be specified. In this case, the predicted value of the DC prediction according to the prior art can be obtained.

(Step S3) The DC prediction unit 203 makes DC prediction by referring to the pixel values of the reference pixels constituting each group. That is, the DC prediction unit 203 generates prediction candidate values for each group. Specifically, the DC prediction unit 203 calculates a statistic for each group with respect to the pixel values of the pixels constituting each group. As the statistic, the average value, the maximum value, the minimum value, the median value, and the like may be applied. For example, when a group is specified by region division based on coordinate values and an average value is applied as a statistic, the average value calculated for each group corresponds to a local average value. Further, for example, when a group is specified by clustering and an average value is applied as a statistic, the average value calculated for each group corresponds to the center of gravity of each cluster. Next, the DC prediction unit 203 proceeds to step S5.

(Step S5) The DC prediction unit 203 selects a prediction value (that is, a selection prediction value) to be applied to coding from the prediction candidate values. In the selection of the selected prediction value, the difference between the image signal of the prediction candidate value and the image signal of the prediction target block P is calculated by, for example, the MSE (Mean Squared Error) method. Then, the DC prediction unit 203 sets the prediction candidate value having the smallest difference as the selection prediction value. The DC prediction unit 203 generates a prediction image based on the selected prediction value, and outputs the image signal of the generated prediction image to the switching unit 111. Further, the DC prediction unit 203 outputs the flag information indicating the selection prediction value to the entropy coding unit 105. After that, the DC prediction unit 203 ends the process shown in FIG.

The flag information output by the DC prediction unit 203 is encoded by the entropy coding unit 105. For example, in the case of the four divisions shown in FIG. 1, the flag information is represented by 2 [bit]. Specifically, "00" is assigned to the first group DC _A1 , "01" is assigned to _{the second group DC A2 , "10" is assigned to the third group DC L1,} and "11" is assigned to the fourth group DC _L2. May be assigned, etc.

Here, as an example, the processing flow of the DC prediction unit 203 in the coding device 10 has been described, but the processing of the DC prediction unit 203 in the decoding device 30 is also the same. However, in the decoding device 30, in step S5, the DC prediction unit 203 selects the predicted value with reference to the flag information decoded by the entropy decoding unit 301.
The above is the description of the operation of the DC prediction unit 203.

As described above, the coding device 10 and the decoding device 30 include a specific unit 204 that specifies a plurality of reference pixel sets (for example, groups) for one image block (for example, a prediction target block). , A generation unit 205 that generates a prediction block candidate by DC prediction from each of a set of a plurality of reference pixels specified by the specific unit 204, and a prediction block for an image block (for example, a prediction composed of pixels of a selection prediction value). An image) is provided with a selection unit 206 for selecting from a plurality of prediction block candidates (for example, a prediction image composed of pixels of prediction candidate values) generated by the generation unit 205.
As a result, the coding device 10 and the decoding device 30 generate candidates for a plurality of prediction blocks by DC prediction for one image block, and select one of them. That is, the coding device 10 and the decoding device 30 can change the prediction block according to the image block. Therefore, the image coding / decoding system 1 can improve the coding efficiency.

Further, the specifying unit 204 specifies a set composed of adjacent reference pixels among the plurality of reference pixels.
As a result, the coding device 10 and the decoding device 30 can perform DC prediction based on the properties of the local image such as the local average value. For example, the coding device 10 and the decoding device 30 can perform DC prediction without impairing the properties of a local image even when the pixel values of the reference pixels are not uniform. Therefore, the image coding / decoding system 1 can improve the coding efficiency.

Further, the specific unit 204 specifies a predetermined number of sets.
As a result, the number of prediction block candidates can be matched between the coding device 10 and the decoding device 30.
In addition, the identification unit 204 specifies a set of reference pixels by clustering.
Thereby, in the coding device 10 and the decoding device 30, the size of the set of reference pixels and the range of the reference pixels included in each set can be changed according to the image block. Therefore, the image coding / decoding system 1 can improve the coding efficiency as compared with the case where the size of the set and the range of the reference pixels included in each set are fixed.

[Second Embodiment]
A second embodiment of the present invention will be described. In addition, here, the same reference numerals are given to the same configurations as those of the first embodiment described above, and the description thereof will be incorporated.
The image coding / decoding system 1A (not shown) according to the present embodiment is an information processing system that encodes and decodes an image in the same manner as the image coding / decoding system 1. However, the image coding / decoding system 1A divides one prediction target block P in DC prediction, and assigns one of a plurality of prediction candidate values to each of the divided blocks. It is different from the conversion / decoding system 1.

Here, the outline of the DC prediction according to the present embodiment will be described.
FIG. 7 is a diagram showing an outline of DC prediction.
In the example shown in FIG. 7, the prediction target block P is equally divided into two in the horizontal direction and the vertical direction, and four divided regions PA, PB, PC, and PD are set. Further, the plurality of reference pixels R are divided into _{four groups, DC A1} , DC _A2 , DC _L1 , and DC _{L2, as in the first embodiment.} Here, in the DC prediction according to the present embodiment, prediction candidate values based on _{the first to fourth groups DC A1} , DC _A2 , DC _L1 , and DC _L2 are used for each of the divided regions PA, PB, PC, and PD. One of them is assigned as a predicted value. As a result, even if the pixel values of the prediction target block P are not uniform, the prediction value can be changed for each division region, so that the prediction accuracy can be improved.
The above is the description of the outline of the DC prediction according to the present embodiment.

Next, the configuration of the image coding / decoding system 1A will be described.
The image coding / decoding system 1A includes a coding device 10A (not shown) and a decoding device 30A (not shown) in place of the coding device 10 and the decoding device 30 included in the image coding / decoding system 1.
The coding device 10A includes an intra prediction unit 200A instead of the intra prediction unit 200 included in the coding device 10. Similarly, the decoding device 30A includes an intra prediction unit 200A instead of the intra prediction unit 200 included in the decoding device 30.

FIG. 8 is a block diagram showing the configuration of the intra prediction unit 200A.
The intra prediction unit 200A includes a DC prediction unit 203A instead of the DC prediction unit 203. The DC prediction unit 203A includes a selection unit 206A instead of the selection unit 206. Further, the DC prediction unit 203A includes a block division unit 207A in addition to each configuration provided in the DC prediction unit 203.

The block division unit 207A divides the prediction target block into a plurality of division areas. The method for dividing the prediction target block by the block division unit 207A is predetermined, and the same prediction target block is similarly divided regardless of whether the coding device 10 or the decoding device 30. The method of allocating the identification information of the divided area is the same. The block division unit 207A outputs the division area information to the selection unit 206A. The divided area information includes, for example, identification information of each divided area and information indicating the range of each divided area.

Similar to the selection unit 206, the selection unit 206A selects one selection prediction value from a plurality of prediction candidate values. However, the selection unit 206A is different from the selection unit 206 in that the block division unit 207A selects the selection prediction value for each division area. Since the method of selecting the selection prediction value by the selection unit 206A is the same as that of the selection unit 206, the description is incorporated. In the case of the coding device 10A, the selection unit 206A outputs flag information indicating the selection prediction value for each division region to the entropy coding unit 105. As a result, the flag information is encoded and transmitted to the decoding device 30A. Further, in the case of the decoding device 30A, the selection unit 206A refers to the flag information decoded by the entropy decoding unit 301 and acquires the selection prediction value for each division region.
The above is the description of the configuration of the image coding / decoding system 1A.

Next, the operation of the image coding / decoding system 1A will be described.
FIG. 9 is a flowchart showing the operation of the DC prediction unit 203A.
Here, as an example, the processing flow of the DC prediction unit 203A in the coding apparatus 10A will be described.
Since the processes of steps S1 and 3 shown in FIG. 9 are the same as the processes shown in FIG. 6, the description is incorporated.
(Step S14) After the process of step S3, the DC prediction unit 203A divides the prediction target block into a plurality of regions. After that, the DC prediction unit 203A proceeds to step S15.
Here, the DC prediction unit 203A divides the prediction target block in either the horizontal direction and the vertical direction, or both. The DC prediction unit 203A may divide the prediction target block into two or more arbitrary numbers. The DC prediction unit 203A may divide the prediction target block evenly in each direction, or may divide the prediction target block unevenly. The DC prediction unit 203A may change the number of divisions in each direction according to the shape of the block to be predicted, for example, the length in each direction. For example, the number of divisions may be increased in the direction of long length and the number of divisions may be decreased in the direction of short length. Further, the DC prediction unit 203A may change the method of dividing the prediction target block according to the method of specifying the group of reference pixels.

(Step S15) The DC prediction unit 203A selects a selection prediction value from the prediction candidate values for each division region of the prediction target block. For example, the DC prediction unit 203A calculates the difference between the predicted image and the original image of each divided region for each prediction candidate value, and sets the prediction candidate value having the smallest difference as the selected prediction value. Other processing is the same as in step S5 shown in FIG. Then, the DC prediction unit 203A ends the process shown in FIG.
The above is the description of the operation of the DC prediction unit 203A.

Next, an example of the processing result by the image coding / decoding system 1A will be described.
FIG. 10 is a diagram showing an example of a processing result by the image coding / decoding system 1A.
FIG. 10A shows an original image of the prediction target block and its DCT signal. FIG. 10B shows a predicted image by DC prediction according to the prior art, a predicted residual, and a DCT signal of the predicted residual. FIG. 10C shows a predicted image by Planar prediction, a predicted residual, and a DCT signal of the predicted residual. FIG. 10D shows a predicted image by the DC prediction unit 203A, a predicted residual, and a DCT signal of the predicted residual. In the example shown in FIG. 10D, the DC prediction unit 203A classifies the reference pixels into eight groups and divides the prediction target block into 16 blocks.

Here, when the residual DCT signals of FIGS. 10B to 10D are compared, the DC component is reduced in the prediction by the DC prediction unit 203A as compared with the DC prediction and the Planar prediction according to the prior art. I understand. After calculating the difference between the original image by MSE, is the DC prediction according to the prior art is "1.49 × 10 ^5". In the Planar prediction is "1.36 × 10 ^5". The prediction by the DC prediction unit 203A is a "1.01 × 10 ^5". As described above, in the prediction by the DC prediction unit 203A, the DC component could be reduced.
The above is an explanation of an example of the processing result by the image coding / decoding system 1A.

As described above, in the coding device 10A and the decoding device 30A, the selection unit 206A selects a prediction block for each of the divided blocks obtained by dividing the image block into a plurality of predicted block candidates from the prediction block candidates.
As a result, the coding device 10A and the decoding device 30A determine the predicted value for each divided block. Therefore, the coding device 10A and the decoding device 30A are, for example, a prediction block that matches the local fluctuation of the pixel value even when the pixel value fluctuates greatly locally in the image block. Can be selected. Therefore, the image coding / decoding system 1A can improve the coding efficiency.

[Modification example]
Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to the above-described embodiments, and includes designs and the like within a range that does not deviate from the gist of the present invention. For example, the configurations described in the first and second embodiments described above can be arbitrarily combined. Further, for example, each configuration described in the above-described first and second embodiments can be omitted when it is not necessary to exert a specific function. Further, for example, each configuration described in the above-described first and second embodiments can be separately provided in a plurality of devices.

In the second embodiment described above, the case where the selected predicted value for each divided region is transmitted as flag information has been described, but the present invention is not limited to this. For example, some selection prediction values may be implicitly defined.

Here, the setting of the implicit selection prediction value will be described.
In a general image, there are few parts where the pixel value changes abruptly. Therefore, even within the prediction target block, it is expected to show the same pixel value as the reference pixel in the region where the distance from the reference pixel is short. Therefore, for example, for the divided region adjacent to the reference pixel, the prediction candidate value of the group including the reference pixel may be implicitly set as the selection prediction value. That is, in the prediction target block, the prediction candidate value based on the group of adjacent reference pixels is selected as the selection prediction value for the divided region located at the end on the reference pixel side. As a result, the coding device 10A and the decoding device 30A can reduce the amount of data while preventing the deterioration of the image quality due to the coding and decoding. That is, the coding efficiency can be improved.

Further, the coding device 10A and the decoding device 30A set the prediction candidate value of the group including the reference pixel in the reference pixel closest to the divided area with respect to the divided area not adjacent to the reference pixel. It may be implicitly set as a selection prediction value. In this case, the distance from the divided area may be, for example, the distance from the representative value with the coordinates of the pixel at the upper left end of the divided area as the representative value.
Further, for example, in each of the horizontal direction and the vertical direction, the group including the reference pixel closest to the distance is specified, and the prediction candidate value by each group is selected by weighted averaging according to the distance ratio to each group. It may be set as a predicted value.
Further, for the divided region not adjacent to the reference pixel, for example, a predicted value used in DC prediction according to the prior art such as Non-Patent Document 1 may be set as a selection candidate value.

Further, a program for realizing the functions of the coding devices 10 and 10A and the decoding devices 30 and 30A described above is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into the computer system. , The processing as the coding devices 10, 10A and the decoding devices 30, 30A may be performed by executing the above. Here, "loading and executing a program recorded on a recording medium into a computer system" includes installing the program in the computer system. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. Further, the "computer system" may include a plurality of computer devices connected via a network including a communication line such as the Internet, WAN, LAN, and a dedicated line. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, or a storage device such as a hard disk built in a computer system. As described above, the recording medium in which the program is stored may be a non-transient recording medium such as a CD-ROM. The recording medium also includes an internal or external recording medium that can be accessed from the distribution server to distribute the program. The code of the program stored in the recording medium of the distribution server may be different from the code of the program in a format that can be executed by the terminal device. That is, the format stored in the distribution server does not matter as long as it can be downloaded from the distribution server and installed in a form that can be executed by the terminal device. The program may be divided into a plurality of parts, downloaded at different timings, and then combined by the terminal device, or the distribution server for distributing each of the divided programs may be different. Furthermore, a "computer-readable recording medium" is a volatile memory (RAM) inside a computer system that serves as a server or client when a program is transmitted via a network, and holds the program for a certain period of time. It shall also include things. Further, the above program may be for realizing a part of the above-mentioned functions. Further, it may be a so-called difference file (difference program) that can realize the above-mentioned function in combination with a program already recorded in the computer system.

1, 1A ... Image coding / decoding system, 10, 10A ... Coding device, 30, 30A ... Decoding device, 50 ... Transmission storage unit, 200, 200A ... Intra prediction unit, 203, 203A ... DC prediction unit, 204 ... Specific Unit, 205 ... Generation unit, 206, 206A ... Selection unit, 207A ... Block division unit, 105 ... Entropy coding unit, 301 ... Entropy decoding unit

Claims (9)

A specific part that specifies a plurality of sets of reference pixels for one image block,
A generation unit that generates a prediction block candidate by intra- prediction from each of the set of the plurality of reference pixels specified by the specific unit.
A selection unit that selects a prediction block for the image block from a plurality of prediction block candidates generated by the generation unit, and a selection unit.
Bei to give a,
The specific unit includes a first group formed by selecting a pixel having the smallest pixel value and a second group formed by selecting a pixel having the largest pixel value among the reference pixels. At least identify as a set,
Predictor. A specific part that specifies a plurality of sets of reference pixels for one image block,
A generation unit that generates a prediction block candidate by intra-prediction from each of the set of the plurality of reference pixels specified by the specific unit.
A selection unit that selects a prediction block for the image block from a plurality of prediction block candidates generated by the generation unit, and a selection unit.
With
The selection unit selects a prediction block for each of the divided blocks obtained by dividing the image block into a plurality of divided blocks from the prediction block candidates .
The prediction block candidate selected by the selection unit is implicitly defined for a part of the division block.
Predictor. The prediction block candidate selected by the selection unit is implicitly defined as the prediction block candidate of the set including the reference pixel for the division block adjacent to the reference pixel.
The prediction device according to claim 2. The prediction block candidate selected by the selection unit is implicitly determined to be the prediction block candidate of the set including the reference pixel closest to the division block for the division block not adjacent to the reference pixel. Has been
The prediction device according to claim 2 or 3. The specific part specifies a predetermined number of sets.
The prediction device according to any one of claims 1 to 4. The specific unit specifies a set of reference pixels by clustering.
The prediction device according to any one of claims 1 to 5. An encoding unit that encodes flag information indicating a prediction block selected by the prediction device according to any one of claims 1 to 6 .
A coding device comprising. Decrypt unit coding apparatus based on the encoded flag information, decoding the image blocks according to claim 7,
Decoding device. A program for operating a computer as a predictor according to any one of claims 1 to 6.

Images