JP2015211269A - Moving image encoding device, moving image encoding method and computer program for moving image encoding - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a moving image encoding device capable of selecting such an orthogonal transform size that image quality deterioration of a decoded picture can be suppressed.SOLUTION: A moving image encoding device 1 includes: a complexity degree calculation section 11 for calculating a complexity degree representing complexity of a scene captured in an encoding target block within a picture; a size selection section 13 by which, regarding each of a plurality of orthogonal transform sizes being different from each other, an evaluation value representing a code amount in performing the orthogonal transform on the encoding target block for the unit of a sub block having that orthogonal transform size is calculated while being weighted so as to be reduced as the complexity degree is reduced and the orthogonal transform size is small, and the orthogonal transform size minimizing the evaluation value is selected; and an encoding section 14 for encoding an orthogonal transform coefficient obtained by performing the orthogonal transform on a prediction error image between the encoding target block and a prediction block for the unit of a sub block having the selected orthogonal transform size.

Description

  The present invention relates to, for example, a moving picture coding apparatus, a moving picture coding method, and a moving picture coding computer program for dividing a picture included in moving picture data into a plurality of blocks and coding each block.

  The moving image data generally has a very large amount of data. Therefore, a device that handles moving image data compresses the moving image data by encoding it when transmitting the moving image data to another device or when storing the moving image data in the storage device. .

  At that time, the picture to be encoded is divided into a plurality of blocks and encoded for each block. A moving image encoding apparatus that encodes moving image data encodes, with respect to an encoding target block, a picture encoded before a picture including the block, or an already encoded image within the picture including the block. A prediction block is generated from the converted block. In order to reduce redundancy by using a high spatial correlation between the encoding target block and the prediction block, the moving image encoding apparatus represents a prediction error image representing an error between the encoding target block and the prediction block. And the orthogonal transform coefficient is obtained by orthogonal transform that reduces redundancy by using the high spatial correlation of the pixels in the block. Then, the moving picture coding apparatus quantizes the orthogonal transform coefficient and then performs variable length coding.

  Furthermore, H.264 MPEG-4 Advanced Video Coding (H.264 MPEG-4 AVC) or H.265 (sometimes called High Efficiency Video Coding (HEVC)), which is one of the video coding systems. The video encoding apparatus can select the size of a block that is a prediction unit or an orthogonal transform unit according to a picture. Therefore, the video encoding apparatus calculates an evaluation value of the encoded information amount for each combination of a plurality of division modes, prediction modes, and transform size modes prepared in advance for the encoding target block. Then, the moving image encoding device encodes the block with a combination that minimizes the evaluation value (see, for example, Patent Document 1).

JP 2007-243427 A

  The scene shown in the picture may be a scene in which a small object moves at high speed in a relatively flat background, such as a scene in which a ball thrown to the ground rolls. In such a case, since there are few characteristic parts in the picture and the amount of movement of the object is large, when the picture is coded by the inter prediction coding mode in which the picture is coded with reference to other pictures, motion prediction is performed. May not be appropriate. That is, as a result of the motion search process for generating the prediction block, an area in which only the background is shown is selected, and the moving object may not be included in the prediction block. In such a scene, even when the encoding target block is encoded in the intra prediction encoding mode in which the encoding target block is encoded using the information of the already encoded area in the picture, the prediction block is It is created from the area where only the background is shown. Therefore, in the prediction error image, the pixels included in the region where the moving object exists have a value other than 0, but the pixels included in the other regions are almost zero. In such a case, the larger the block size, the smaller the evaluation value of the coding amount. Therefore, a relatively large block is selected as the orthogonal transform block size. As the size of the orthogonal transform block increases, the ratio of the area including the small moving object to the block decreases. Therefore, the orthogonal transform coefficient resulting from the area is a high-frequency component having a relatively small absolute value. As a result, the quantized orthogonal transform coefficient resulting from the region including the moving object by the quantization process becomes 0, and information about the moving object is lost. There is no moving object. Thus, if an appropriate size is not selected for a block that is a unit of orthogonal transform, the image quality of the decoded picture is degraded.

  Accordingly, an object of the present specification is to provide a moving picture coding apparatus capable of selecting an orthogonal transform size that can suppress degradation in image quality of a decoded picture.

  According to one embodiment, a moving image encoding apparatus that encodes a picture included in moving image data is provided. The moving image encoding apparatus includes: a dividing unit that divides a picture into a plurality of blocks; a complexity calculating unit that calculates complexity representing the complexity of a scene shown in an encoding target block among the plurality of blocks; A prediction block generation unit that generates a prediction block of a coding target block from a coding target picture including a coding target block or a picture encoded before the coding target picture, and a plurality of orthogonal transforms different from each other For each of the sizes, the evaluation value indicating the code amount when the encoding target block is orthogonally transformed in units of sub-blocks having the orthogonal transformation size is such that the lower the complexity, the smaller the evaluation value becomes. The size selection unit that selects the orthogonal transform size that minimizes the evaluation value, the encoding target block, and the prediction block And an encoding unit for encoding the orthogonal transform coefficients obtained by orthogonal transformation in units of sub-blocks having the selected orthogonal transformation size prediction error image between click.

The objects and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims.
It should be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are not restrictive of the invention as claimed.

  The moving image encoding apparatus disclosed in this specification can select an orthogonal transform size that can suppress degradation in image quality of a decoded picture.

It is a figure explaining the influence on the image quality of the decoded picture by the size of the block which applies orthogonal transformation. It is a schematic block diagram of the moving image encoder which concerns on one embodiment. It is a figure which shows the relationship between the complexity of an encoding object block, and a weighting coefficient. It is an operation | movement flowchart of a moving image encoding process. It is a figure which shows the relationship between the complexity of a subblock, and a weighting coefficient by a modification. It is a figure which shows an example of the picture division | segmentation by HEVC. It is a figure which shows an example of the process at the time of determining the application size of orthogonal transformation based on HEVC. It is a block diagram of the computer which operate | moves as a moving image encoder by the computer program which implement | achieves the function of each part of the moving image encoder by embodiment or its modification.

  Hereinafter, the moving picture coding apparatus will be described with reference to the drawings. First, the influence of the block size to which the orthogonal transform in the encoding target block is applied will be described.

  FIG. 1 is a diagram for explaining the influence on the image quality of a decoded picture due to the size of a block to which orthogonal transform is applied. A prediction error image 100 having a size of 16 × 16 pixels includes a region 101 representing a moving object. This region 101 is small, for example, a size less than 4 × 4 pixels. It is assumed that the prediction error image 100 is divided into sub-blocks each having a size of 4 × 4 pixels, which is an application unit of orthogonal transformation, and one of the sub-blocks 110 includes the region 101. In this case, when each sub-block is orthogonally transformed, the size of the sub-block 110 is small. Therefore, in the sub-block 110, the orthogonal transformation coefficient becomes a relatively large value due to the influence of the region 101. Note that, in the sub-blocks other than the sub-block 110, the prediction error image 100 is flat, so that the orthogonal transform coefficient is almost 0 except for the DC component. Since the orthogonal transform coefficient has a relatively large value, as shown in the sub-block 120 in which the orthogonal transform coefficient of the sub-block 110 is quantized, in the sub-block 120, the quantized orthogonal transform coefficient other than the DC component is used. At least some of them have non-zero values. As a result, the area 101 is also reproduced in the decoded picture 130.

  On the other hand, when the entire prediction error image 100 is orthogonally transformed as one block 140, since the region 101 is relatively small with respect to the block 140, the absolute value of the orthogonal transformation coefficient is also small. As a result, as shown in the block 150 in which the orthogonal transform coefficient is quantized, the quantized orthogonal transform coefficient becomes 0 for the components other than the direct current due to the quantization process, and the information in the region 101 is lost. To do. As a result, the area 101 disappears in the decoded picture 160.

  Such degradation of the picture quality of a decoded picture due to the disappearance of a small area is difficult for the viewer to notice when the scene shown in the picture is relatively complex, but when the background is relatively flat. It will stand out.

  Therefore, in the present embodiment, the moving image encoding apparatus calculates a complexity representing the degree of complexity of a scene shown in the block to be encoded. Then, when determining the size of a block that is an application unit of orthogonal transform, the video encoding apparatus uses a weighting factor that decreases as complexity decreases, and a code amount of a term that increases as the orthogonal transform size decreases. , And an evaluation value of the code amount of the encoding target block is calculated. As a result, the moving picture coding apparatus makes it easy to select a relatively small size as a block size to which the orthogonal transformation is applied for a block in which a flat scene is captured.

  Note that the picture may be either a frame or a field. The frame is one still image in the moving image data, while the field is a still image obtained by extracting only odd-numbered data or even-numbered data from the frame.

FIG. 2 is a schematic configuration diagram of a moving image encoding apparatus according to an embodiment. The moving image encoding device 1 includes a dividing unit 10, a complexity calculating unit 11, a prediction block generating unit 12, a size mode selecting unit 13, and an encoding unit 14. The encoding unit 14 includes a prediction error calculation unit 15, an orthogonal transform unit 16, a quantization unit 17, a decoding unit 18, a storage unit 19, and a variable length encoding unit 20.
Each of these units included in the moving image encoding apparatus 1 is formed as a separate circuit. Alternatively, these units included in the video encoding device 1 may be mounted on the video encoding device 1 as one integrated circuit in which circuits corresponding to the respective units are integrated. Furthermore, each of these units included in the moving image encoding device 1 may be a functional module realized by a computer program executed on a processor included in the moving image encoding device 1.

A control unit (not shown) that controls the entire moving image coding apparatus 1 sequentially converts pictures included in moving image data into a dividing unit 10 according to the coding order specified by, for example, Group Of Pictures (GOP). Is input. A GOP includes a plurality of consecutive pictures and represents a structure in which an encoding method for each picture is defined.
The dividing unit 10 divides a picture to be encoded into a plurality of blocks having a predetermined number of pixels. Each block has a maximum size among applicable sizes of orthogonal transform, for example, a size of 16 horizontal pixels × 16 vertical pixels. Then, the dividing unit 10 outputs each block in the raster scan order, for example.

  The complexity calculation unit 11 calculates a complexity representing the degree of complexity of the scene shown in the encoding target block among the blocks output from the dividing unit 10. This complexity is used to determine the weighting coefficient when calculating the evaluation value of the coding amount when determining the size of the block to be orthogonally transformed and the prediction mode. Note that the prediction mode defines a method for generating a prediction block, and represents an inter prediction encoding mode or an intra prediction encoding mode. Furthermore, the prediction mode represents an intra prediction mode that defines a method for generating a prediction block when the intra prediction coding mode is selected, and a unidirectional prediction mode when the inter prediction coding mode is selected. Or bi-directional prediction mode.

ここでpiは、ブロック内の画素の値であり、mは、ブロック内の全画素の平均画素値である。Mは、ブロック内に含まれる画素の総数であり、例えば、ブロックが16×16画素のサイズを持つ場合、M=256である。そしてActはブロックのアクティビティである。 In the present embodiment, the complexity calculation unit 11 calculates the sum (called an activity) of the average pixel value of all the pixels in the block and the absolute difference value of each pixel. The activity is calculated by the following formula.

Here, pi is a value of a pixel in the block, and m is an average pixel value of all the pixels in the block. M is the total number of pixels included in the block. For example, when the block has a size of 16 × 16 pixels, M = 256. Act is a block activity.

あるいは、複雑度算出部１１は、ブロック内の画素値の分散を、複雑度として算出してもよい。 The complexity calculation unit 11 may calculate the activity PixAct per pixel as the complexity according to the following equation instead of the block activity.

Alternatively, the complexity calculation unit 11 may calculate the variance of the pixel values in the block as the complexity.

  The complexity calculation unit 11 outputs the complexity to the size mode selection unit 13.

  The prediction block generation unit 12 generates a prediction block for each prediction mode applicable to the encoding target block.

  For example, if the encoding target picture is an I picture, the prediction block generating unit 12 encodes an encoded block adjacent to the left side or the upper side of the encoding target block for each intra prediction mode that defines a prediction block generation method. Generate a prediction block from The encoding target picture is a picture including the encoding target block. The I picture is a picture that is a target of intra prediction encoding and is not subjected to inter prediction encoding.

  Also, if the picture to be encoded is a P picture, the prediction block generation unit 12 performs a motion search between a reference picture obtained by decoding an encoded picture and the block to be encoded, for example. Then, the prediction block generation unit 12 calculates a motion vector indicating a region on the reference picture that most closely matches the encoding target block, and sets the region indicated by the motion vector as a prediction block. Also, if the encoding target picture is a B picture, the prediction block generation unit 12 performs motion search in both the pictures preceding and following the encoding target picture in two directions, for example, in the order of playback time, An area that most closely matches the encoding target block on two pictures is obtained. And the prediction block production | generation part 12 calculates | requires the average value of the value of the pixel in the corresponding position in these two area | regions, and makes it a prediction block. Furthermore, for the P picture or B picture, the prediction block generation unit 12 may generate a prediction block for each intra prediction mode, as in the case where the encoding target picture is an I picture. Further, the prediction block generation unit 12 may generate a prediction block for the B picture by unidirectional motion prediction, similarly to the P picture. Note that a P picture is a picture that can be inter-predictively encoded using information of one already encoded picture. A B picture is a picture that can be bidirectionally inter-predictively encoded using information of two already encoded pictures. Each block in the P picture may be intra prediction encoded. Also, each block in the B picture may be intra prediction encoded or may be inter prediction encoded in one direction.

  When the prediction block generation unit 12 can further change the size of the block to which the prediction mode is applied or the size of the block to be subjected to motion search, the prediction block generation unit 12 can also change each of the plurality of subblocks obtained by dividing the encoding target block. In addition, a prediction block may be generated.

  The prediction block generation unit 12 outputs the prediction block generated for the encoding target block and the prediction mode and motion vector corresponding to the prediction block to the size mode selection unit 13.

ここで、Dは、予測ブロックと符号化対象ブロック間の差分演算により算出される予測誤差画像の符号量の評価値である予測誤差画像評価値である。Rは、直交変換の適用サイズに応じて算出される、符号化項目情報量である。またαは、複雑度に応じて決定される重み係数であり、λは定数である。 The size mode selection unit 13 is an example of a size selection unit, and determines a block size and a prediction mode, which are units of orthogonal transform applied to an encoding target block. For example, the orthogonal transform block size is selected from 16 × 16 pixels, 8 × 8 pixels, and 4 × 4 pixels. Therefore, the size mode selection unit 13 evaluates the estimated value of the code amount of the encoding target block for each block size serving as an orthogonal transform unit and for each prediction block generated by the prediction block generation unit 12. E is calculated. At that time, the size mode selection unit 13 calculates the evaluation value by weighting so that the evaluation value becomes smaller as the complexity becomes lower and the smaller the orthogonal transform size becomes.

Here, D is a prediction error image evaluation value that is an evaluation value of the code amount of the prediction error image calculated by the difference calculation between the prediction block and the encoding target block. R is an encoded item information amount calculated according to the application size of orthogonal transform. Α is a weighting factor determined according to complexity, and λ is a constant.

The size mode selection unit 13 divides the prediction error image into N subblocks according to the application size of the orthogonal transform in order to obtain the prediction error image evaluation value D for each application size of the orthogonal transform. To Hadamard transform. For example, if the application size of orthogonal transformation is 8 × 8 pixels, N = 4, and if the application size of orthogonal transformation is 4 × 4 pixels, N = 16. The size mode selection unit 13 sums the absolute values of the coefficients C _ji (i = 1, 2,..., M, where M is the number of pixels included in the sub-block) obtained by Hadamard transform SATD _j (j = 1,2, ..., N) is calculated for each sub-block. Then, the size mode selection unit 13 sets the sum of the absolute sum SATD _j of the Hadamard transform coefficients of each sub-block as the prediction error image evaluation value D.

  Further, the size mode selection unit 13 calculates, as the encoding item information amount R, code amounts such as an orthogonal transform size division flag, a quantization value applied to the encoding target block, and a coefficient presence / absence flag. Note that the orthogonal transform size division flag represents an application size of orthogonal transform. The coefficient presence / absence flag represents the presence / absence of an orthogonal transformation coefficient whose absolute value is not 0, which is obtained for each of the case where the pixel value is luminance and the color difference, and is obtained for each sub-block subjected to orthogonal transformation. Therefore, since the number of coefficient presence / absence flags increases as the block size for orthogonal transform decreases, the encoded item information amount R increases as the block size for orthogonal transform decreases.

  The weighting factor α is set to be smaller as the complexity of the encoding target block is lower. The weighting coefficient α is used to weight the encoded item information amount R that increases as the block size of the orthogonal transform becomes smaller when the evaluation value E is calculated, as shown in the equation (3).

  FIG. 3 is a diagram illustrating the relationship between the complexity of the block to be encoded and the weighting factor. In FIG. 3, the horizontal axis represents complexity (in this example, activity per pixel), and the vertical axis represents the value of the weighting factor α. The graph 300 represents the relationship between the complexity and the weighting factor α. In this example, when the complexity is 0, that is, when all the pixels included in the encoding target block have the same value, the weighting coefficient α is a minimum of 0.2. As the complexity increases, the weighting factor α also increases linearly. When the complexity is 10 or more, the weighting factor α is also constant at 1.

  By setting the weighting factor α in this way, the value of the term including the encoded item information amount R becomes smaller as the scene shown in the encoding target block becomes flatter. For this reason, the flatter the scene shown in the encoding target block, the easier it is to select a smaller size as the application size of the orthogonal transform.

The size mode selection unit 13 selects the orthogonal transform size and the prediction mode corresponding to the minimum evaluation value as being used for encoding the block to be encoded. Therefore, the size mode selection unit 13 outputs the prediction block generated according to the selected prediction mode and the orthogonal transform size to the encoding unit 14.
Further, the size mode selection unit 13 also outputs a parameter representing the selected prediction mode and information used to generate a prediction block in the selected prediction mode, for example, a motion vector, to the encoding unit 14. .

  The encoding unit 14 encodes the encoding target block according to the prediction block generated according to the prediction mode selected by the size mode selection unit 13 and the orthogonal transform size. That is, the encoding unit 14 encodes an orthogonal transform coefficient obtained by orthogonal transform of a prediction error image between a coding target block and a prediction block in units of sub blocks having a selected orthogonal transform size.

  Therefore, the prediction error calculation unit 15 of the encoding unit 14 performs a difference calculation between the encoding target block and the prediction block. And the prediction error calculation part 15 makes the difference value corresponding to each pixel in the block obtained by the difference calculation a prediction error image. The prediction error calculation unit 15 outputs the prediction error image and the orthogonal transform size to the orthogonal transform unit 16 of the encoding unit 14.

  The orthogonal transform unit 16 divides the prediction error image of the encoding target block into sub blocks having the selected orthogonal transform size, and performs orthogonal transform in units of sub blocks, thereby obtaining orthogonal transform coefficients for each sub block. For example, the orthogonal transform unit 16 obtains a set of DCT coefficients for each sub-block as orthogonal transform coefficients by performing discrete cosine transform (DCT) on each sub-block as orthogonal transform processing. . However, when the selected orthogonal transform size and the size of the encoding target block are the same, the orthogonal transform unit 16 may perform orthogonal transform processing on the encoding target block itself.

  The orthogonal transform unit 16 outputs the orthogonal transform coefficient of each sub-block of the encoding target block to the quantization unit 17 of the encoding unit 14.

  The quantization unit 17 calculates the quantization coefficient of the orthogonal transform coefficient by quantizing each orthogonal transform coefficient of the encoding target block. This quantization process is a process that represents a signal value included in a certain section as one signal value. The fixed interval is called a quantization width. For example, the quantization unit 17 quantizes the orthogonal transform coefficient by truncating a predetermined number of lower bits corresponding to the quantization width from the orthogonal transform coefficient. The quantization width is determined by the quantization parameter. For example, the quantization unit 17 determines a quantization width to be used according to a function representing a quantization width value with respect to a quantization parameter value. The function can be a monotonically increasing function with respect to the value of the quantization parameter, and is set in advance. Also, the quantization parameter is determined by, for example, a control unit (not shown) based on the code amount allocated to the encoding target picture including the encoding target block, and is notified to the quantization unit 17.

  The quantization unit 17 outputs the quantization coefficient of the block to be encoded to the decoding unit 18 and the variable length coding unit 20.

  The decoding unit 18 generates a reference block and a reference picture for encoding a block after the block from the quantization coefficient of the encoding target block. For this purpose, the decoding unit 18 inversely quantizes the quantization coefficient by multiplying the quantization coefficient by a predetermined number corresponding to the quantization width determined by the quantization parameter. By this inverse quantization, an orthogonal transform coefficient of the encoding target block, for example, a set of DCT coefficients is restored. Thereafter, the decoding unit 18 performs an inverse orthogonal transform process on the orthogonal transform coefficient for each sub-block having the size of the applied orthogonal transform. For example, when the orthogonal transform unit 16 calculates an orthogonal transform coefficient using DCT, the decoding unit 18 performs an inverse DCT process on the restored orthogonal transform coefficient. By performing the inverse quantization process and the inverse orthogonal transform process on the quantized signal, a prediction error image having the same level of information as the prediction error image before encoding is reproduced.

The decoding unit 18 adds the reproduced prediction error signal corresponding to the pixel to each pixel value of the prediction block. By executing these processes for each block, the decoding unit 18 generates a reference block used to generate a prediction block for a block to be encoded thereafter. The decoding unit 18 may perform a deblocking filter process on the reference block in order to reduce block noise of the reference block.
Each time the decoding unit 18 generates a reference block, the decoding unit 18 stores the reference block in the storage unit 19.

  The storage unit 19 temporarily stores the reference block received from the decoding unit 18. A reference picture is obtained by combining reference blocks for one picture in accordance with the coding order of each block. The storage unit 19 supplies the reference picture or the reference block to the prediction block generation unit 12 and the size mode selection unit 13. The storage unit 19 stores a predetermined number of reference pictures that may be referred to by the encoding target picture. If the number of reference pictures exceeds the predetermined number, the encoding order is old. Discard in order from the reference picture.

  The variable length coding unit 20 performs variable length coding so that the quantization coefficient received from the quantization unit 17 is shorter as the signal value has a higher occurrence probability. The variable length coding unit 20 also performs variable length coding on information used for generating a prediction block such as a motion vector. The variable length coding unit 20 can use, for example, Huffman coding processing such as CAVLC or arithmetic coding processing such as CABAC as the variable length coding processing.

  The moving image encoding apparatus 1 adds predetermined information including a prediction mode for each block as header information to the encoded signal generated by the variable length encoding unit 20, thereby encoding the moving image. A data stream including image data is generated. The moving image encoding apparatus 1 stores the data stream in a storage unit (not shown) having a magnetic storage medium, an optical storage medium, a semiconductor memory, or the like, or outputs the data stream to another device.

  FIG. 4 is an operation flowchart of the moving image encoding process executed by the moving image encoding device 1. The moving image encoding device 1 executes the moving image encoding process shown in FIG. 4 for each encoding target block.

  The complexity calculator 11 calculates the complexity of the encoding target block (step 101). Then, the complexity calculation unit 11 outputs the complexity to the size mode selection unit 13. Further, the prediction block generation unit 12 generates a prediction block for each prediction mode applicable to the encoding target block (step S102). The prediction block generation unit 12 outputs the prediction block to the size mode selection unit 13.

  The size mode selection unit 13 determines the weighting factor α so that the weighting factor α with respect to the encoded item information amount R decreases as the complexity decreases (step S103). Then, the size mode selection unit 13 calculates the evaluation value E by weighting the encoding item information amount R by the weighting coefficient α for each prediction mode applicable to the encoding target block and applicable block size of orthogonal transform. (Step S104). The size mode selection unit 13 sets the prediction mode and orthogonal transform block size that minimize the evaluation value E as the prediction mode and orthogonal transform block size to be applied to the encoding target block (step S105). Then, the size mode selection unit 13 outputs the prediction block generated in the prediction mode to be applied and the block size of the orthogonal transform to the encoding unit 14.

  The prediction error calculation unit 15 of the encoding unit 14 performs a difference calculation between the encoding target block and the prediction block to calculate a prediction error image (step S106). Then, the orthogonal transform unit 16 of the encoding unit 14 divides the prediction error image in units of sub blocks having the applied orthogonal transform size, and performs orthogonal transform for each sub block to calculate orthogonal transform coefficients (step S107). .

  The quantizing unit 17 of the encoding unit 14 obtains a quantized coefficient by quantizing the orthogonal transform coefficient (step S108). Then, the quantization unit 17 outputs the quantized coefficients to the decoding unit 18 and the variable length coding unit 20 of the coding unit 14.

  The decoding unit 18 dequantizes the quantized coefficient to reproduce the orthogonal transform coefficient, and applies the inverse orthogonal transform to the orthogonal transform coefficient for each sub-block having the block size applied in the orthogonal transform. Reproduce the prediction error image. And the decoding part 18 adds the value of each pixel of a prediction error image to the value of the corresponding pixel of a prediction block, calculates | requires a reference block, and memorize | stores the reference block in the memory | storage part 19 (step S109).

  On the other hand, the variable length coding unit 20 performs variable length coding on each quantization coefficient included in the coding target block (step S110). Then, the moving image encoding apparatus 1 ends the moving image encoding process for the encoding target block.

  Note that the moving image encoding apparatus 1 may exchange the order of the process of step S101 and the process of S102, or may execute the process of step S101 and the process of S102 in parallel.

  As described above, this moving image encoding apparatus is adapted to the encoding item information amount as the complexity of the encoding target block is lower, that is, as the scene shown in the encoding target block is flatter. Reduce the weighting factor. Therefore, the smaller the complexity of the block to be encoded, the easier it is to select a smaller size as the application size for orthogonal transform. As a result, this moving image encoding apparatus can suppress an image quality deterioration in which a small moving object disappears in a decoded image even in a scene where a small moving object is reflected with a flat background. Can be selected.

  According to the modification, the complexity calculation unit 11 may calculate the complexity from the prediction block generated for the encoding target block. In this case, when a plurality of prediction blocks are generated by the prediction block generation unit 12, the complexity calculation unit 11 calculates the complexity for each prediction block, and calculates the average value for the complexity of the encoding target block. It is good.

  Alternatively, the complexity calculating unit 11 selects, as the encoding target block, a prediction block in which the total sum of the absolute values of the pixels of the prediction error image between the encoding target block and the prediction block is the smallest among the prediction blocks. Estimated to be most similar. Therefore, the complexity calculation unit 11 may calculate the complexity from a prediction block that minimizes the sum of the absolute values of the pixels of the prediction error image.

ここで、SATD_jは、符号化対象ブロックの予測誤差画像をN個のサブブロックに分割したときのj番目のサブブロックについて算出される予測誤差評価値であり、アダマール変換係数C_jiの絶対値の総和として算出される。Rは符号化項目情報量であり、λは定数である。またβ_jは、j番目のサブブロックについて算出された複雑度に応じて決定される重み係数である。 According to another modification, the complexity calculating unit 11 may divide the encoding target block into sub-blocks that are units of orthogonal transform, and calculate the complexity for each sub-block.
In this case, it is preferable that the size mode selection unit 13 calculates the evaluation value E according to the following equation instead of the equation (3).

Here, SATD _j is a prediction error evaluation value calculated for the j-th sub-block when the prediction error image of the encoding target block is divided into N sub-blocks, and the absolute value of the Hadamard transform coefficient C _ji Is calculated as the sum of R is an encoded item information amount, and λ is a constant. Β _j is a weighting factor determined according to the complexity calculated for the j-th sub-block.

In this modification, the weighting factor β _j is set to increase as the complexity decreases.
FIG. 5 is a diagram showing the relationship between the complexity of sub-blocks and the weighting factor according to this modification. In FIG. 5, the horizontal axis represents complexity (in this example, activity per pixel), and the vertical axis represents the value of the weighting factor β _j . The graph 500 represents the relationship between the complexity and the weighting coefficient β _j . In this example, when the complexity is 0, that is, when all the pixels included in the encoding target block have the same value, the weight coefficient β _j is 1 at the maximum. As the complexity increases, the weight coefficient β _j decreases linearly. When the complexity is 10 or more, the weight coefficient β _j is constant at 0.2.

The smaller the size of the sub-block, the higher the complexity with a slight variation in pixel values within the sub-block. In particular, when the scene shown in the encoding target block is a scene in which a small object is moving in a flat background, the larger the subblock including the area representing the moving object, the larger the subblock. Complexity decreases. Therefore, by setting the weighting factor β _j as described above, the weighting factor for the prediction error evaluation value for the subblock including the region representing the moving object becomes smaller as the size of the subblock becomes smaller. As a result, a small size is easily selected as an application size of the orthogonal transformation in the encoding target block in which the scene as described above is reflected.

  According to another modification, the size mode selection unit 13 may determine the application size of the prediction mode and the orthogonal transform so as to comply with HEVC.

  FIG. 6 is a diagram illustrating an example of picture division by HEVC. As shown in FIG. 6, a picture 600 is divided in coding block coding tree unit (CTU) units, and each CTU 601 is encoded in the raster scan order. The size of the CTU 601 can be selected from 64 × 64 to 16 × 16 pixels.

  The CTU 601 is further divided into a plurality of Coding Units (CU) 602 in a quadtree structure. Each CU 602 in one CTU 601 is encoded in the Z scan order. The size of the CU 602 is variable, and the size is selected from CU division modes 8 × 8 to 64 × 64 pixels. The CU 602 is a unit for selecting an intra prediction encoding mode and an inter prediction encoding mode that are encoding modes. The CU 602 is individually processed in units of Prediction Unit (PU) 603 or Transform Unit (TU) 604. The PU 603 is a unit for performing prediction according to the encoding mode. For example, the PU 603 is a unit to which the intra prediction mode is applied in the intra prediction coding mode, and a unit for performing motion compensation in the inter prediction coding mode. The size of the PU 603 can be selected, for example, from the PU partition modes PartMode = 2Nx2N, NxN, 2NxN, Nx2N, 2NxU, 2NxnD, nRx2N, and nLx2N in inter prediction encoding. On the other hand, the TU 604 is a unit of orthogonal transformation, and the size of the TU 604 is selected from 4 × 4 pixels to 32 × 32 pixels. The TU 604 is divided by a quadtree structure and processed in the Z scan order.

  In this case, the complexity calculation unit 11 calculates the complexity, for example, with the CU as an encoding target block. Then, the complexity calculation unit 11 outputs the complexity calculated for the CU to the size mode selection unit 13.

  The size mode selection unit 13 calculates an evaluation value E for each of the encoding target block and the encoding target block when the encoding target block is divided into four sub blocks. For example, it is assumed that the size of the CU that is the encoding target block is 32 × 32 pixels. In this case, the size mode selection unit 13 evaluates each of the evaluation values in a case where the CU itself is an application unit of orthogonal transformation and a case where orthogonal transformation is performed for each 16 × 16 pixel sub-block obtained by equally dividing the CU into four. E is calculated. When there are a plurality of applicable prediction modes, the size mode selection unit 13 calculates an evaluation value E for each prediction mode. The size mode selection unit 13 uses the evaluation value when the encoding target block itself is the application unit of orthogonal transform, and the evaluation value when the sub-block obtained by dividing the encoding target block into four is the application unit of orthogonal transformation. Is smaller than the encoding target block itself as an application unit of orthogonal transform. On the other hand, when the evaluation value when the sub-block obtained by dividing the block to be encoded into four sub-blocks is set as the application unit of the orthogonal transform is smaller than the case where the encoding target block itself is set as the application unit of the orthogonal transform, The unit 13 sets each sub block as an encoding target block. Then, the size mode selection unit 13 calculates an evaluation value and selects an orthogonal transform size to be applied to each encoding target block. At that time, the complexity calculation unit 11 also calculates the complexity for the newly set encoding target block.

  The size mode selection unit 13 repeats the above processing until the encoding target block itself becomes an application unit of orthogonal transformation or the size of the sub-block becomes 4 × 4 pixels which is the minimum setting value of the TU size.

  FIG. 7 is a diagram illustrating an example of processing when determining the application size of orthogonal transform based on HEVC. In this example, the size mode selection unit 13 calculates the evaluation value E according to equation (3). However, the complexity mode calculation unit 11 calculates the complexity for each of the sub-blocks obtained by dividing the block or sub-block that is the target of determination of whether or not to divide into four into four, so that the size mode selection unit 13 ( 4) The evaluation value E may be calculated according to the equation.

  First, the complexity calculator 11 and the size mode selector 13 use the CU 700 as an encoding target block, the CU 700 itself as a TU, and the 16 × 16 pixel sub-block obtained by dividing the CU 700 into four TUs. The evaluation value E is calculated respectively. If the evaluation value when the CU 700 itself is TU is smaller than the evaluation value when the 16 × 16 pixel sub-block obtained by dividing the CU 700 into TUs, the TU 710 has the same size as the CU 700.

  On the other hand, when the evaluation value when the TU is a sub-block of 16 × 16 pixels obtained by dividing the CU 700 into four is smaller than the evaluation value when the CU 700 itself is set as the TU, the complexity calculation unit 11 calculates the 16 × 16 The complexity is calculated for each sub-block 701 of pixels. Further, the size mode selection unit 13 calculates an evaluation value when each subblock 701 is a TU, and an evaluation value when an 8 × 8 pixel subblock obtained by dividing each subblock into four is a TU.

When the evaluation value when each sub-block 701 is set to TU is smaller than the evaluation value when the sub-block of 8 × 8 pixels obtained by dividing each sub-block 701 into four is set as TU, the size mode selection unit 13 , Each sub-block 701 is designated as TU 711.
On the other hand, for any of the sub-blocks 701, for example, the upper-right sub-block 701a, the evaluation when the sub-block 701a is divided into four TUs is used as the TU rather than the evaluation value when the sub-block 701a is set as the TU. Suppose the value is smaller. In this case, the complexity calculation unit 11 calculates the complexity for each 8 × 8 pixel sub-block 702 obtained by dividing the sub-block 701a into four. Further, the size mode selection unit 13 calculates an evaluation value when each subblock 702 is set as a TU, and an evaluation value when a subblock of 4 × 4 pixels obtained by dividing each subblock 702 into four is set as a TU. .

When the evaluation value when each sub-block 702 is TU is smaller than the evaluation value when the sub-block of 4 × 4 pixels obtained by dividing each sub-block 702 into four is set as TU, the size mode selection unit 13 Each sub-block 702 in the sub-block 701a is set as TU712. Further, for subblocks 701 other than the subblock 701a, the subblock 701 is designated as TU711.
On the other hand, for any of the subblocks 702, for example, the lower left subblock 702a, the evaluation when the subblock 702a is divided into four TUs is used as the TU rather than the evaluation value when the subblock 702a is set as the TU. Suppose the value is smaller. In this case, the size mode selection unit 13 sets each subblock in the subblock 702a as TU713. For subblocks 702 other than the subblock 702a, the subblock 702 is designated as TU712.

  According to this modification, the moving image coding apparatus can select an appropriate orthogonal transform size that can suppress image quality degradation and can also select an orthogonal transform size based on HEVC, as in the above-described embodiment. .

  FIG. 8 is a configuration diagram of a computer that operates as a moving image encoding apparatus when a computer program that realizes the functions of the respective units of the moving image encoding apparatus according to the above-described embodiment or its modification is operated.

  The computer 100 includes a user interface unit 101, a communication interface unit 102, a storage unit 103, a storage medium access device 104, and a processor 105. The processor 105 is connected to the user interface unit 101, the communication interface unit 102, the storage unit 103, and the storage medium access device 104 via, for example, a bus.

  The user interface unit 101 includes, for example, an input device such as a keyboard and a mouse, and a display device such as a liquid crystal display. Alternatively, the user interface unit 101 may include a device such as a touch panel display in which an input device and a display device are integrated. Then, the user interface unit 101 outputs, for example, an operation signal for selecting moving image data to be encoded to the processor 105 in accordance with a user operation.

  The communication interface unit 102 may include a communication interface for connecting the computer 100 to a device that generates moving image data, for example, a video camera, and a control circuit thereof. Such a communication interface can be, for example, Universal Serial Bus (Universal Serial Bus, USB).

  Furthermore, the communication interface unit 102 may include a communication interface for connecting to a communication network according to a communication standard such as Ethernet (registered trademark) and a control circuit thereof.

  In this case, the communication interface unit 102 acquires moving image data to be encoded from another device connected to the communication network, and passes the data to the processor 105. Further, the communication interface unit 102 may output the encoded moving image data received from the processor 105 to another device via a communication network.

  The storage unit 103 includes, for example, a readable / writable semiconductor memory and a read-only semiconductor memory. The storage unit 103 stores a computer program for executing a moving image encoding process executed on the processor 105, and data generated during or as a result of these processes.

  The storage medium access device 104 is a device that accesses a storage medium 106 such as a magnetic disk, a semiconductor memory card, and an optical storage medium. For example, the storage medium access device 104 reads a computer program for moving image encoding processing executed on the processor 105 stored in the storage medium 106 and passes the computer program to the processor 105.

  The processor 105 generates encoded moving image data by executing the computer program for moving image encoding processing according to the above-described embodiment or modification. The processor 105 stores the generated encoded moving image data in the storage unit 103 or outputs it to another device via the communication interface unit 102.

  Note that the computer program capable of executing the functions of the respective units of the moving image encoding device 1 on the processor may be provided in a form recorded on a computer-readable medium. However, such a storage medium does not include a carrier wave.

  All examples and specific terms listed herein are intended for instructional purposes to help the reader understand the concepts contributed by the inventor to the present invention and the promotion of the technology. It should be construed that it is not limited to the construction of any example herein, such specific examples and conditions, with respect to showing the superiority and inferiority of the present invention. Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions and modifications can be made thereto without departing from the spirit and scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Moving image encoder 10 Divider 11 Complexity calculator 12 Prediction block generator 13 Size mode selector 14 Encoder 15 Prediction error calculator 16 Orthogonal transformer 17 Quantizer 18 Decoder 19 Storage unit 20 Variable length Encoding unit 100 Computer 101 User interface unit 102 Communication interface unit 103 Storage unit 104 Storage medium access device 105 Processor 106 Storage medium

Claims (8)

A moving image encoding apparatus for encoding a picture included in moving image data,
A dividing unit for dividing the picture into a plurality of blocks;
A complexity calculator that calculates the complexity representing the complexity of the scene shown in the encoding target block of the plurality of blocks;
A prediction block generation unit that generates a prediction block of the encoding target block from an encoding target picture including the encoding target block, or a picture encoded before the encoding target picture;
For each of a plurality of different orthogonal transform sizes, the evaluation value indicating the code amount when the encoding target block is orthogonally transformed in units of sub-blocks having the orthogonal transform size is reduced as the complexity decreases. A size selection unit that calculates weighted so that the evaluation value becomes smaller as the transform size, and selects an orthogonal transform size that minimizes the evaluation value;
An encoding unit that encodes an orthogonal transform coefficient obtained by orthogonal transform of a prediction error image between the encoding target block and the prediction block in units of sub-blocks having the selected orthogonal transform size;
A moving picture encoding apparatus having: The evaluation value includes a first term that increases as the orthogonal transform size decreases.
The moving image encoding apparatus according to claim 1, wherein the size selection unit calculates the evaluation value by weighting the first term with a weighting factor that decreases as the complexity decreases.   The video code according to claim 2, wherein the term includes information defined for each of the plurality of sub-blocks when the encoding target block is divided into sub-blocks having the orthogonal transform size. Device.   4. The moving picture encoding apparatus according to claim 3, wherein the information defined for each of the plurality of sub-blocks includes a flag indicating whether or not the sub-block has the orthogonal transform coefficient whose absolute value is not 0. 5. . The evaluation value includes a second term representing a code amount for each subblock when the encoding target block is divided into subblocks having the orthogonal transform size,
The complexity calculation unit calculates the complexity for each sub-block when the encoding target block is divided into sub-blocks having the orthogonal transform size;
The size selection unit calculates the evaluation value by weighting the second term with a weighting factor that increases as the complexity calculated for the subblock decreases for each subblock. Video encoding device. The size selection unit includes a first evaluation value that is the evaluation value when the size of the encoding target block is the orthogonal transform size, the encoding target block into four first sub-blocks, etc. The second evaluation value, which is the evaluation value when the size of the first sub-block when divided is the orthogonal transform size, is calculated, and the first evaluation value is the second evaluation value. When smaller than the evaluation value, the size of the encoding target block is selected as the orthogonal transform size,
On the other hand, when the second evaluation value is smaller than the first evaluation value, the complexity calculation unit sets the four first sub-blocks as the encoding target blocks to the complexity calculation unit. And calculating each of the four first sub-blocks as the coding target block, calculating the first evaluation value and the second evaluation value, and calculating the first evaluation value and the second evaluation value. The moving image encoding apparatus according to claim 1, wherein the orthogonal transform size is selected according to a comparison result of second evaluation values. A moving image encoding apparatus for encoding a picture included in moving image data,
Dividing the picture into a plurality of blocks;
Calculating the complexity representing the complexity of the scene shown in the encoding target block of the plurality of blocks;
Generating a prediction block of the encoding target block from an encoding target picture including the encoding target block or a picture encoded before the encoding target picture;
For each of a plurality of different orthogonal transform sizes, the evaluation value indicating the code amount when the encoding target block is orthogonally transformed in units of sub-blocks having the orthogonal transform size is reduced as the complexity decreases. Calculate by weighting so that the evaluation value becomes smaller as the transform size, select the orthogonal transform size that minimizes the evaluation value,
Encoding an orthogonal transform coefficient obtained by orthogonally transforming a prediction error image between the encoding target block and the prediction block in units of subblocks having the selected orthogonal transform size;
A moving picture encoding method including the above. A moving image encoding computer program for encoding a picture included in moving image data,
Dividing the picture into a plurality of blocks;
Calculating the complexity representing the complexity of the scene shown in the encoding target block of the plurality of blocks;
Generating a prediction block of the encoding target block from an encoding target picture including the encoding target block or a picture encoded before the encoding target picture;
For each of a plurality of different orthogonal transform sizes, the evaluation value indicating the code amount when the encoding target block is orthogonally transformed in units of sub-blocks having the orthogonal transform size is reduced as the complexity decreases. Calculate by weighting so that the evaluation value becomes smaller as the transform size, select the orthogonal transform size that minimizes the evaluation value,
Encoding an orthogonal transform coefficient obtained by orthogonally transforming a prediction error image between the encoding target block and the prediction block in units of subblocks having the selected orthogonal transform size;
A computer program that causes a computer to execute the operation.

Images