JP5846838B2 - Moving picture coding method, moving picture coding apparatus, moving picture decoding method, moving picture decoding apparatus, and programs thereof - Google Patents

Description

  The present invention reduces the code amount of the correction value and reduces the generated code amount in the moving image coding in the moving image coding in which the correction value is given to the pixel group classified by the predetermined rule by the in-loop filter. The present invention relates to an encoding / decoding technique for performing the above.

  In the present specification, a structure that groups pixel groups classified according to a predetermined rule is defined as a class. The rules for classifying pixel groups are explained in the “Background” section below.

  The next standardization standard for moving picture coding, which is being standardized as High Efficiency Video Coding (HEVC), In addition to the deblocking filter mounted on H.264, a Wiener filter-based adaptive filter and Sample Adaptive Offset (SAO) [see Non-Patent Document 1] for correcting the amplitude direction of pixel values are added.

  An outline of SAO processing will be described. In the present specification, nine types designated by the syntax element pqao_type_idx shown in FIG. 18 are defined as SAO types. In SAO, first, each pixel is classified into a plurality of classes based on the characteristics of the pixel value of the decoded image. Subsequently, the screen is divided by a hierarchical expression based on the quadtree structure, and correction to the corresponding pixels classified for each class is performed on the area corresponding to the leaf node of the quadtree.

  There are two categories of classes: band offset (BO) and edge offset (EO).

(1) band offset (BO)
FIG. 19 shows the class classification in BO. Based on the range of pixel values (0 to 255 for 8-bit signals), the range is divided into 32 at equal intervals. Each section is defined as a class in BO, and a correction value is calculated for each class. The 32 divided sections are divided into two groups of 16 sections, and only one group designated as shown in FIG. 19 transmits the correction value. Here, the first group in FIG. 19 is BO_1 and the second group is BO_2.

(2) edge offset (EO)
Class classification is performed based on the magnitude relationship with neighboring pixel values. As a classification method of magnitude relation, four kinds of methods based on the magnitude relation with 2 neighboring pixels and two kinds of methods based on the magnitude relation with 4 neighboring pixels are prepared.

  FIG. 20 is a diagram illustrating a method of selecting neighboring pixels in EO, and FIG. 21 is a diagram illustrating class classification in EO. As shown in FIG. 21, the specific class classification includes five classes (class classification for FIGS. 20A to 20D) in the method based on the magnitude relationship with two neighboring pixels, and the magnitude relation with four neighboring pixels. In the method based on this, the data is classified into seven classes (class classifications for FIGS. 20E and 20F), and a correction value is calculated for each class. Here, the four types based on the magnitude relationship with the two neighboring pixels are EO_1 to 4, respectively, and the methods based on the magnitude relationship with the four neighboring pixels are EO_5 and 6, respectively.

  Each correction value is stored in the slice header of the encoded stream. The correction values for 16 classes are respectively stored in BO_1 and 2; 5 classes are respectively stored in EO_1 to 4; and 7 classes are respectively stored in EO_5 and 6; Decrypted.

  At the time of encoding, any one of the methods that do not apply BO_1, 2 or EO_1 to 6 or SAO to the region corresponding to the leaf node of the quadtree is selected by RD optimization, and a number that clearly indicates the selected method (see FIG. 18) is added to the encoded stream as the SAO type.

  At the time of decoding, for the area corresponding to the leaf node of the quadtree, a number (SAO type) specifying the method selected at the time of encoding is decoded from the encoded stream, and the pixel value is corrected by applying the method.

Chih-Ming Fu, Ching-Yeh Chen, Yu-Wen Huang and Shawmin Lei, “CE8 Subset3: Picture Quadtree Adaptive Offset”, JCTVC-D122, January, 2011

  For example, in an image with a dark entire screen such as an image in a black fade process, the pixels assigned to each class of BO are concentrated in a class having a low pixel value, and the number of assigned pixels in a class having a high pixel value is almost zero. As for EO, a class in which the number of assigned pixels is zero may occur depending on the pattern in the screen.

  However, in the conventional SAO, as described above, all correction values for the corresponding class are encoded / decoded. That is, some correction value is transmitted as additional information even for a class in which the number of assigned pixels is zero, leaving room for improvement in encoding efficiency.

  An object of the present invention is to solve the above-described problems, reduce the amount of codes generated for additional information in SAO, and improve encoding efficiency.

  In the present invention, the number of assigned pixels of each class is analyzed in advance with respect to the decoded image at the time of encoding, and when the correction value given to each class is encoded, the number of assigned pixels is non-zero. Only the correction value of the class is encoded and added to the encoded stream, and for the class with the number of assigned pixels of zero, an adaptive process that omits the process of storing the correction value in the encoded stream is executed.

  At the time of decoding, before decoding the correction value given to each class, the number of assigned pixels of each class is analyzed in advance with respect to the decoded image. Only a correction value of a class having a non-zero number of assigned pixels is decoded, and an adaptive process that omits decoding of the correction value is executed for a class having a zero number of assigned pixels.

  Note that the number of assigned pixels for each class can be calculated in the decoding apparatus using a processing procedure common to the encoding apparatus. Therefore, based on the number of assigned pixels for each class, the decoding device can determine whether or not it is necessary to decode the correction value of the class, and it is not necessary to add encoding information for this determination. Thereby, it is possible to reduce the generated code amount of the correction value in SAO.

  According to the present invention, it is possible to reduce SAO additional information in the in-loop filter, and as a result, it is possible to reduce the amount of generated code in moving picture coding.

It is a figure which shows one structural example of the moving image encoder to which this invention is applied. It is a figure which shows the structural example of the in-loop filter process part in a moving image encoder. It is a flowchart of the in-loop filter process in a moving image encoding. It is a figure which shows the structural example of the SAO process part in a moving image encoder. It is a flowchart of the SAO process in a moving image encoding. It is a flowchart of the SAO process in a moving image encoding. It is a figure which shows the structural example of the correction value encoding part in a moving image encoder. It is a flowchart of the correction value encoding process in moving image encoding. It is a figure which shows one structural example of the moving image decoding apparatus to which this invention is applied. It is a figure which shows the structural example of the in-loop filter process part in a moving image decoding apparatus. It is a flowchart of the in-loop filter process in a moving image decoding. It is a figure which shows the structural example of the SAO process part in a moving image decoding apparatus. It is a flowchart of the SAO process in moving image decoding. It is a figure which shows the structural example of the correction value decoding part in a moving image decoding apparatus. It is a flowchart of the correction value decoding process in moving image decoding. It is a figure which shows the hardware structural example in the case of comprising a moving image encoder by a computer and a software program. It is a figure which shows the hardware structural example in the case of comprising a moving image decoding apparatus with a computer and a software program. It is a figure which shows the type of SAO designated by the syntax element pqao_type_idx. It is a figure which shows the class classification in BO. It is a figure which shows the selection method of the neighborhood pixel in EO. It is a figure which shows the class classification in EO.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[Configuration example of moving image encoding apparatus]
FIG. 1 is a diagram illustrating a configuration example of a moving image encoding apparatus to which the present invention is applied. In the moving image encoding apparatus 100, in the present embodiment, the part of the in-loop filter processing unit 110 is particularly different from the prior art, and the other parts are the same as the configuration of the conventional general moving image encoding apparatus. It is.

  The video encoding apparatus 100 receives a video signal to be encoded, divides a frame of the input video signal into blocks, encodes each block, and outputs the bit stream as an encoded stream. The prediction residual signal generation unit 103 obtains a difference between the block of the input video signal and the prediction block of the intra prediction unit 101 or the inter prediction unit 102, and uses it as a prediction residual signal. The transform unit 104 performs orthogonal transform (DCT) on the prediction residual signal and outputs transform coefficients. The quantization unit 105 quantizes the transform coefficient and outputs the quantized transform coefficient.

  In the inverse quantization unit 106, the quantization unit 105 performs inverse quantization, and in the inverse transform unit 107, transform coefficients are inversely transformed. The inverse transformed value is added and synthesized with the prediction block by the decoded signal generation unit 108 and stored in the decoded image storage unit 109. The image in the decoded image storage unit 109 is used in the intra prediction unit 101. When the above processing is completed for all the blocks of the frame, the in-loop filter processing unit 110 applies an image processing filter for removing the coding distortion of the decoded image, and the filtered output signal is stored in the frame buffer 111. Is done.

  The image in the frame buffer 111 is used by the inter prediction unit 102. Prediction information and quantized transform coefficients of the intra prediction unit 101 and the inter prediction unit 102 and the overhead of the in-loop filter processing unit 110 are input to the entropy coding unit 112. The entropy encoding unit 112 performs entropy encoding on the input information and outputs an encoded stream.

[Configuration Example of In-Loop Filter Processing Unit in Video Encoding Device]
FIG. 2 is a diagram illustrating a configuration example of an in-loop filter processing unit in a moving image encoding apparatus to which the present invention is applied. In the in-loop filter processing unit 110, in the present embodiment, a part of the SAO processing unit 1102 is particularly different from the prior art, and the other parts are the same as the configuration of a conventional general video encoding apparatus.

  The in-loop filter processing unit 110 receives the signal before filtering processing and the original signal as input, the deblocking filter processing unit 1101 reduces block noise, the SAO processing unit 1102 corrects the pixel value in the amplitude direction, The adaptive filter processing unit 1103 performs a total of three stages of filter processing to minimize the square error with the original signal. The filtered signal is stored in the frame buffer 111 and referred to by inter prediction. The SAO processing unit 1102 and the adaptive filter processing unit 1103 each output various coefficients, and these outputs are input to the entropy encoding unit 112 and stored in the encoded stream.

[Example of in-loop filter processing procedure in video coding]
The function of the in-loop filter processing unit 110 described above is to remove the coding noise of the decoded signal by the three filter processes included. The input of the in-loop filter process is the decoded signal before noise removal and the original signal, and the output is the decoded signal after noise removal and the overhead of the in-loop filter.

FIG. 3 is a flowchart of in-loop filter processing in moving image coding. Hereinafter, processing in each step shown in FIG. 3 will be described.
Step S101: Read an original signal and a decoded signal.
Step S102: A deblocking filter is executed to reduce block noise.
Step S103: SAO is executed to correct the pixel value in the amplitude direction, and various coefficients such as a filter coefficient, a block size, and a flag indicating application / non-application of the filter are determined.
Step S104: An adaptive filter is executed to perform a filter process that minimizes a square error with the original signal, and various coefficients such as a filter coefficient, a block size, and a flag that indicates whether the filter is not applied are determined.
Step S105: The various coefficients of SAO and the various coefficients of the adaptive filter are output as the in-loop filter overhead, and the decoded signal after noise removal is output.

[Configuration Example of SAO Processing Unit in Video Encoding Device]
The SAO processing unit 1102 shown in FIG. 2 determines the optimum SAO type and applicable block size while subdividing the block with a quadtree structure from the maximum block of a predetermined size.

  FIG. 4 shows a detailed configuration example of the SAO processing unit that executes SAO processing on an arbitrary block. In the present embodiment, the correction value encoding unit 204 is particularly different from the prior art, and the other parts are the same as the configuration of a conventional general video encoding apparatus. In the method of the present invention, a decoded signal is given to the correction value encoding unit 204 as an additional input.

  First, the optimum SAO type / correction value determination unit 201 determines the SAO type and correction value that minimize the RD cost among all types of pqao_type_idx (0 to 8) shown in FIG. 18 for the target block. , Generate the target block after correction. The division determination unit 202 determines whether or not the target block is divided into four square shapes based on the size of the target block.

  The quadtree division flag / SAO type encoding unit 203 encodes the division flag and the optimal SAO type, and the correction value encoding unit 204 corrects the optimum SAO type / correction value determination unit 201 to determine the correction. After encoding the value, the RD cost calculation unit 205 calculates the RD cost from the generated code amount and the corrected target block.

  If it is determined by the division determination unit 202 that the target block can be divided, the quad-tree division flag encoding unit 206 encodes the division flag, and the first division block calculation unit 207 to the fourth division block calculation are performed. Each of the units 210 repeats recursive processing by the processing unit shown in FIG. 4 (FIG. 4), and outputs the RD cost, the corrected divided block, and various coefficients of each divided block.

  The division determination unit 211 determines the RD cost of the target block and the total value of the four RD costs of the divided blocks, and outputs the RD cost, the corrected block, and various coefficients with the lower RD cost. If it is determined by the division determination unit 202 that the target block cannot be divided, the quadtree division flag / SAO type encoding unit 203 omits the division flag encoding process, and the quadtree division flag encoding unit 206. A sufficiently large value is substituted for the total value of the RD costs of the four divided blocks calculated by the first divided block calculation unit 207 to the fourth divided block calculation unit 210.

[Example of SAO processing procedure in video encoding]
The function of the SAO processing unit is to calculate the SAO type and correction value of each target block divided into a quadtree structure. The input of the SAO processing unit is the decoded signal before correction, the original signal, the coordinate value of the target block, and the size of the target block, and the output is the target block after correction, the RD cost, and various coefficients.

5 and 6 are flowcharts of SAO processing in moving image coding. The processing performed by the SAO processing unit shown in FIG. 4 according to the steps shown in FIGS. 5 and 6 will be described.
Step S201: Read the decoded signal before correction, the coordinate value of the target block, the size of the target block, and the original signal.
Step S202: The optimum SAO type and correction value of the target block are calculated, and a corrected target block is generated.
Step S203: Judgment of whether or not division is possible is performed from the coordinate value and size of the target block.
Step S204: Perform conditional branching according to the possibility of division based on the determination result of step S203.
Step S205: When division is possible, first, for the case where division is possible but not division, division flag (off) encoding and optimum SAO type encoding are performed, and the process proceeds to step S207.
Step S206: When division is impossible, optimal SAO type encoding is performed.
Step S207: Encode the correction value of each class used in the optimum SAO type.
Step S208: Calculate the RD cost of the target block.
Step S209: Perform conditional branching on whether or not division is possible based on the determination result in step S203. If division is impossible, the process proceeds to step S215, and if possible, the following processing is executed.
Step S210: The RD cost × 4 is initialized to 0, and the division flag (on) is encoded as a divided block stream.
Step S211: The process up to Step S214 is repeated for each divided block divided into a square shape (loop L1).
Step S212: This process is recursively called for each of the divided blocks, and the processes after step S201 are performed.
Step S213: The RD cost of the divided block is added to RD cost × 4.
Step S214: When the processing of the loop L1 for each divided block is completed, the process proceeds to step S216.
Step S215: When block division is impossible, a sufficiently large value is substituted for RD cost × 4.
Step S216: Compare the RD cost of the target block with the RD cost × 4, and output the RD cost with the lower cost, the corrected block, and various coefficients.

[Configuration Example of Correction Value Encoding Unit of SAO in Video Encoding Device]
The SAO correction value encoding unit 204 shown in FIG. 4 is a different part between the conventional method and the present embodiment. FIG. 7 shows an example of the configuration of the correction value encoding unit. In particular, FIG. 7A shows an example according to the conventional method, and FIG. 7B shows an example according to the embodiment of the present invention.

  First, a configuration example of the SAO correction value encoding unit 204 ′ in the conventional method will be described with reference to FIG.

  In the case of the conventional correction value encoding unit 204 ′, the encoding target class determination unit 2041 ′ determines each class identified from the optimum SAO type, and the class correction value encoding unit 2042 ′ determines all the targets. Encode class correction values.

  On the other hand, the correction value encoding unit 204 according to the present embodiment is configured as shown in FIG. The encoding target class determination unit 2041 has the same function as the conventional encoding target class determination unit 2041 ′. The intra-class pixel count calculation unit 2043 counts the number of pixels allocated to each SAO type class optimal in the decoded signal. The class correction value encoding unit 2042 encodes a correction value only for a class in which the number of pixels assigned to each class is non-zero.

[SAO correction value encoding processing in moving image encoding]
The function of the correction value encoding unit 204 described above is to encode the correction value of each class corresponding to the optimum SAO type. The input of the correction value encoding unit 204 for that purpose is the optimum SAO type and correction value of each class. In this embodiment, a decoded signal is also input. The output of the correction value encoding unit 204 is various coefficients (encoded correction values for each class).

  FIG. 8 shows a processing flow of the correction value encoding unit, in particular, FIG. 8A shows an example according to the conventional method, and FIG. 8B shows an example according to the embodiment of the present invention.

First, the processing procedure of the SAO correction value encoding unit 204 ′ in the conventional method will be described with reference to FIG.
Step S311: Read the optimum SAO type and correction value for each class.
Step S312: The class to be encoded is identified from the optimum SAO type.
Step S313: Encode correction values for all classes to be processed.

In contrast, the SAO correction value encoding unit 204 according to the present embodiment executes the processing shown in FIG.
Step S301: First, an optimum SAO type, a correction value for each class, and a decoded signal are read.
Step S302: Identify the class to be encoded from the optimum SAO type.
Step S303: Count the number of pixels allocated to each class to be encoded with the optimum SAO type in the decoded signal.
Step S304: In the target class, only the correction value of the class in which the number of pixels allocated in step S303 is non-zero is encoded.

[Configuration example of video decoding device]
FIG. 9 is a diagram illustrating a configuration example of a moving image decoding apparatus to which the present invention is applied. In the moving picture decoding apparatus 300, in the present embodiment, the part of the in-loop filter processing unit 308 is different from the conventional technique, and the other parts are the same as the configuration of the conventional general moving picture decoding apparatus.

  The moving picture decoding apparatus 300 receives an encoded stream, and the entropy decoding unit 301 decodes prediction information, quantized transform coefficients, and in-loop filter overhead for each block. The decoded post-quantization transform coefficients are inversely quantized by the inverse quantization unit 304, and inverse orthogonal transform (IDCT) is performed by the inverse transform unit 305. The output signal of the inverse transform unit 305 is added and synthesized with the prediction block of the intra prediction unit 302 or the inter prediction unit 303 by the decoded signal generation unit 306 and stored in the decoded image storage unit 307.

  The image in the decoded image storage unit 307 is used in the intra prediction unit 302. When the above processing is completed for all the blocks of the frame, an in-loop filter processing unit 308 applies an image processing filter for removing the coding distortion of the decoded image, and stores the filtered output signal in the frame buffer 310. Is done. The image in the frame buffer 310 is used in the inter prediction unit 303.

[Configuration Example of In-Loop Filter Processing Unit in Video Decoding Device]
FIG. 10 is a diagram illustrating a configuration example of an in-loop filter processing unit in the video decoding device to which the present invention is applied. In the in-loop filter processing unit 308, in the present embodiment, the SAO processing unit 3082 is particularly different from the prior art, and the other parts are the same as the configuration of a conventional general video decoding apparatus.

  The in-loop filter processing unit 308 receives the pre-filtering signal and various coefficients, the deblocking filter processing unit 3081 reduces block noise, the SAO processing unit 3082 corrects the pixel value in the amplitude direction, The adaptive filter processing unit 3083 performs a total of three stages of filter processing to minimize the square error with the original signal.

  The filtered signal is stored in the frame buffer 310 and is referenced by the inter prediction unit 303. The SAO processing unit 3082 and the adaptive filter processing unit 3083 each receive various coefficients, and use appropriate filter coefficients and block sizes determined by the moving image coding apparatus, and flags indicating application / non-application of filters. Perform filtering.

[Example of in-loop filter processing procedure in video decoding]
The function of the in-loop filter processing unit 308 described above is to remove the coding noise of the decoded signal by the three included filter processes. The input of the in-loop filter process is the decoded signal before noise removal and the overhead of the in-loop filter, and the output is the decoded signal after noise removal.

FIG. 11 is a flowchart of in-loop filter processing in moving picture decoding. Hereinafter, processing in each step shown in FIG. 11 will be described.
Step S401: Read decoded signal and in-loop filter overhead.
Step S402: A deblocking filter is executed to reduce block noise.
Step S403: The overhead is read and SAO is executed to correct the amplitude direction of the pixel value.
Step S404: The overhead is read and an adaptive filter is executed.

[Configuration Example of SAO Processing Unit in Video Decoding Device]
The above-described SAO processing unit 3082 shown in FIG. 10 performs correction based on the decoded SAO type for each block subdivided into a quadtree structure from the maximum block of a predetermined size.

  FIG. 12 shows a configuration example of a SAO processing unit that executes SAO processing on an arbitrary block. In this embodiment, the correction value decoding unit 404 is particularly different from the conventional technology, and the other parts are the same as the configuration of a conventional general video decoding device. In the method of the present invention, a decoded signal is given to the correction value decoding unit 404 as an additional input.

  First, the division determination unit 401 determines whether or not division flag decoding is necessary from the size of the target block. When the quadtree division flag decoding unit 402 performs decoding, if division is possible, the division flag decoding process is performed. If this is not possible, the decryption process is omitted. At this time, if the division flag is off, the SAO type decoding unit 403 performs processing, the correction value decoding unit 404 decodes each correction value of the SAO type target class, and the SAO execution unit 405 performs SAO. Execute and output the corrected block. When the division flag is on, this figure (FIG. 12) is shown for each divided block obtained by dividing the target block into a square shape by each of the first divided block calculation unit 406 to the fourth divided block calculation unit 409. A recursive calculation process is performed by the processing unit described in 1.

[Example of SAO processing procedure in video decoding]
The function of the SAO processing unit in moving image decoding is to perform correction by SAO for each target block divided into a quadtree structure. The input of the SAO processing unit is a decoded signal before correction, the coordinate value of the target block, the size of the target block, and various coefficients, and the output is the target block after correction.

FIG. 13 is a flowchart of SAO processing in moving picture decoding. Processing performed by the SAO processing unit 3082 shown in FIG. 10 according to the steps shown in FIG. 13 will be described.
Step S501: Read the decoded signal before correction, the coordinate value of the target block, the size of the target block, and various coefficients.
Step S502: It is determined whether or not division is possible based on the size of the target block.
Step S503: A conditional branch indicating whether or not division is possible is performed based on the determination result in step S502. If division is impossible, the process proceeds to step S506.
Step S504: If division is possible, the division flag is decoded.
Step S505: Conditional branching of the division flag is performed. If it is on, the process proceeds to step S509. If it is off, the process proceeds to step S506.
Step S506: Decrypt the SAO type.
Step S507: The correction value in each class of SAO type is decoded.
Step S508: Correct the pixel value of the target block using the correction value and the SAO type, output the corrected block, and end the SAO processing.
Step S509: When the division flag is on, the processing up to step S511 is repeated for each divided block divided into a square shape (loop L1).
Step S510: This process is recursively called for each of the divided blocks, and the processes after Step S501 are performed.
Step S511: When the process of the loop L1 for each divided block is completed, the SAO process is terminated.

[Configuration Example of SAO Correction Value Decoding Unit in Moving Picture Decoding Device]
The aforementioned SAO correction value decoding unit 404 shown in FIG. 12 is different between the conventional method and the present embodiment. FIG. 14 shows a configuration example of the correction value decoding unit, in particular, FIG. 14 (A) shows an example according to the conventional method, and FIG. 14 (B) shows an example according to the embodiment of the present invention.

  The SAO correction value decoding unit 404 ′ in the conventional method is configured as shown in FIG. 14A, and the class correction value decoding unit 4041 ′ decodes the correction values of all the target classes from the optimum SAO type. Do.

  On the other hand, the SAO correction value decoding unit 404 according to the present embodiment is configured as shown in FIG. 14B, and the intra-class pixel count calculation unit 4041 includes each of the optimum SAO type in the decoded signal. Count the number of pixels assigned to the class. The class correction value decoding unit 4042 decodes the correction value only for a class in which the number of pixels assigned to each class is non-zero.

[SAO correction value decoding processing in moving image decoding]
The function of the correction value decoding unit 404 is to decode the correction value of each class corresponding to the optimum SAO type. The input of the correction value decoding unit 404 for that purpose is the optimum SAO type and various coefficients (encoded correction values of each class). In this embodiment, a decoded signal is also input. The output of the correction value decoding unit 404 is a correction value of each class corresponding to the optimum SAO type.

  FIG. 15 shows a processing flow of the correction value decoding unit, in particular, FIG. 15A shows an example according to the conventional method, and FIG. 15B shows an example according to the embodiment of the present invention.

First, the processing procedure of the SAO correction value decoding unit 404 ′ in the conventional method will be described with reference to FIG.
Step S611: The optimum SAO type and various coefficients are read.
Step S612: The correction values for all the classes to be processed are decoded.
Step S613: Output the correction value of the decoded class.

On the other hand, the SAO correction value decoding unit 404 according to the present embodiment executes the processing shown in FIG.
Step S601: Read an optimum SAO type, various coefficients (correction values), and a decoded signal.
Step S602: Count the number of pixels assigned to each class to be decoded of the optimum SAO type in the decoded signal.
Step S603: Only the correction value of the class in which the number of pixels allocated in step S602 is non-zero in the target class is decoded.
Step S604: The decoded class correction value is output.

  The video encoding and video decoding processes described above can be realized by a computer and a software program, and the program can be recorded on a computer-readable recording medium or provided through a network. It is.

  FIG. 16 shows a hardware configuration example in the case where the moving picture encoding apparatus of FIG. 1 is configured by a computer and a software program. This system includes a CPU 700 that executes a program, a memory 701 such as a RAM that stores programs and data accessed by the CPU 700, and a video signal input unit 702 that inputs a video signal to be encoded from a camera or the like (disk device). A program storage device 703 in which a moving image encoding program 704, which is a software program for causing the CPU 700 to execute processing for encoding an input video signal by this method, An encoded stream output unit 705 that outputs encoded data generated by the CPU 700 executing the moving image encoding program 704 loaded in the memory 701 via, for example, a network (stores an encoded stream by a disk device or the like). May be a storage unit) In has become connected to each other.

  FIG. 17 shows a hardware configuration example in the case where the moving picture decoding apparatus of FIG. 9 is configured by a computer and a software program. This system includes a CPU 800 that executes a program, a memory 801 such as a RAM that stores programs and data accessed by the CPU 800, and a code that inputs an encoded stream that is encoded and output by the moving image encoding apparatus according to this method. Stored in an encoded stream input unit 802 (which may be a storage unit that stores an encoded stream by a disk device or the like) and a moving image decoding program 804 that is a software program that causes the CPU 800 to execute a process of decoding the input encoded stream by this method. And the decoded video data output unit 805 that outputs the decoded video obtained by decoding the encoded stream to the playback device by the CPU 800 executing the moving picture decoding program 804 loaded in the memory 801. Are connected by a bus. It has become.

  The embodiments of the present invention have been described above with reference to the drawings. However, the above embodiments are merely examples of the present invention, and it is obvious that the present invention is not limited to the above embodiments. is there. Accordingly, additions, omissions, substitutions, and other modifications of the components may be made without departing from the spirit and technical scope of the present invention.

DESCRIPTION OF SYMBOLS 100 Moving image encoder 110 In-loop filter process part 1102 SAO process part 204 Correction value encoding part 2041 Encoding object class determination part 2042 Class correction value encoding part 2043 Corresponding pixel count calculation part 300 Moving image decoding apparatus 308 In-loop filter processing unit 3082 SAO processing unit 404 Correction value decoding unit 4041 Corresponding number of pixels in class calculation unit 4042 Class correction value decoding unit

Claims (6)

For the decoded image generated in the moving image encoding process, each pixel is classified into a plurality of classes based on the characteristics of the pixel value of the decoded image, pixel correction values are given for each class, and based on the correction values In a moving image encoding method using an in-loop filter having a processing function for correcting a pixel value of a decoded image,
Counting the number of pixels in the decoded signal assigned to each class;
A video encoding method comprising: encoding only a correction value of a class having a non-zero number of pixels allocated to a target class based on the pixel count result and adding the correction value to the encoded stream. . For the decoded image generated in the moving image encoding process, each pixel is classified into a plurality of classes based on the characteristics of the pixel value of the decoded image, pixel correction values are given for each class, and based on the correction values In a moving image encoding apparatus using an in-loop filter having a processing function for correcting a pixel value of a decoded image,
Means for counting the number of pixels in the decoded signal assigned to each class;
A moving picture encoding apparatus comprising: a means for encoding only a correction value of a class having a non-zero number of pixels assigned to a target class based on the pixel count result and adding the correction value to an encoded stream. . For the decoded image generated in the moving image decoding process, the pixel class classified based on the characteristic of the pixel value of the decoded image is identified, and the decoding is performed based on the correction value of the pixel given to the identified class. In a moving picture decoding method using an in-loop filter having a processing function for correcting a pixel value of an image,
Counting the number of pixels in the decoded signal assigned to each class;
And a step of decoding only the correction value of the class in which the number of pixels assigned to the target class is non-zero based on the result of counting the number of pixels. For the decoded image generated in the moving image decoding process, the pixel class classified based on the characteristic of the pixel value of the decoded image is identified, and the decoding is performed based on the correction value of the pixel given to the identified class. In a moving picture decoding apparatus using an in-loop filter having a processing function for correcting a pixel value of an image,
Means for counting the number of pixels in the decoded signal assigned to each class;
A moving picture decoding apparatus comprising: means for decoding only a correction value of a class in which the number of pixels allocated to a target class is non-zero based on the result of counting the number of pixels.   A moving picture coding program for causing a computer to execute the moving picture coding method according to claim 1.   A moving picture decoding program for causing a computer to execute the moving picture decoding method according to claim 3.

Images