JP2010035025A - Encoding device and method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform encoding which achieves easy packaging and efficient in-frame prediction. <P>SOLUTION: A frame buffer 10 stores an image generated by performing local decoding on a predictively encoded image, and a reference prediction mode memory 102 stores information showing a used prediction mode. An SAD (Sum of Absolute Differences) calculation part 107 calculates SADs of prediction errors of an image to be predicted for respective prediction modes. On the basis of information on the prediction mode used for predictive encoding of an image adjacent to the image to be predicted, a syntax length calculation part 108 calculates a syntax length of the image to be predicted, and a probability calculation part 109 calculates the probability that the prediction mode is selected for the predictive encoding of the image to be predicted. An evaluated value calculation part 110 calculates evaluated values for the respective prediction modes from the SADs, syntax length, probability, and a quantization scale QP. A selection part 111 selects a prediction mode based upon the evaluated values. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to coding of a moving image for performing intra-frame prediction.

  Recent dramatic advances in digital signal processing technology have facilitated the recording of moving images to storage media and the transmission of moving images via transmission lines, which were difficult in the past. When such recording and transmission are performed, compression coding processing is performed on each picture constituting the moving image to greatly reduce the data amount of the coded stream. H.264 / AVC is standardized as a moving image compression encoding method.

  In H.264 / AVC, a slice obtained by dividing a picture by a plurality of macroblocks (MB), which are division units, is a basic unit of encoding. One picture is composed of at least one of P (predictive) slice, B (bi-predictive) slice, I (intra) slice, and special P, I slice SP, SI slice for stream switching. The The MB of each slice supports an encoding mode using intra prediction in which prediction encoding is performed using pixels and information in a frame, and inter prediction in which prediction encoding is performed using pixels and information between frames.

  According to H.264 / AVC, there are 4 × 4 pixel, 8 × 8 pixel, and 16 × 16 pixel prediction units for intra prediction of luminance information, and 8 × 8 pixels for intra prediction of chroma information. There is a prediction unit. The intra prediction mode of 4 × 4 pixels is “Intra 4 × 4”, the intra prediction mode of 8 × 8 pixels is “Intra 8 × 8”, the intra prediction mode of 16 × 16 pixels is “Intra 16 × 16”, 8 × The intra prediction mode for chroma information of 8 pixels is called “intra chroma”. Nine prediction formulas corresponding to nine types of prediction modes for intra 4 × 4, nine prediction formulas corresponding to nine types of prediction modes for intra 8 × 8, and four types of prediction for intra 16 × 16 Four prediction formulas corresponding to the mode are defined respectively. In addition, four prediction value arithmetic expressions corresponding to four types of prediction modes are defined for the intra chroma.

  Depending on the prediction target image, intra 4 × 4, intra 8 × 8 or intra 16 × 16 is selected for intra prediction of luminance information, in other words, a prediction unit (image size) is selected. Further, one prediction mode corresponding to the selected image size is selected. Similarly, one prediction mode is selected for intra prediction of chroma information.

  A number indicating the prediction value calculation formula is given to the prediction value calculation formula in each intra prediction mode of luminance information and chroma information. In the standardization stage of H.264, numbers are assigned in descending order of occurrence frequency.

  The prediction value calculation formula when intra 16 × 16 is selected in the intra prediction of luminance information is represented by a 5-bit syntax (mb_type) as shown in FIG.

  On the other hand, when intra 4 × 4 is selected in the intra prediction of luminance information, the syntax of one 4 × 4 block is expressed by 1 bit or 4 bits, and each of 16 images per 1 macroblock (MB) Is set.

  The syntax of the intra 4 × 4 prediction value calculation formula indicates whether or not the prediction value calculation formula number selected in the encoded 4 × 4 block adjacent to the top or the left matches the smaller value. Expressed with 1 bit. In the case of mismatch, the number of the predicted value calculation expression is further indicated by 3 bits. At this time, if it is smaller than the number of the minimum predicted value calculation formula selected in the adjacent encoded 4 × 4 block, it is expressed by the number of the selected predicted value calculation formula. In addition, when it is larger than the number of the minimum predicted value calculation formula selected in the adjacent encoded 4 × 4 block, it is expressed by a number obtained by subtracting 1 from the number of the selected predicted value calculation formula.

  However, H.264 does not standardize the prediction mode determination method. Therefore, JM (Joint Model), which is the basic software of H.264, implements a sum of absolute differences (SAD) method and a rate distortion optimization (RDO) method as selection methods. The SAD method is used when the external input parameter RDOptimization = “0” of JM, and the RDO method is used when RDOptimization = “1”.

  FIG. 1 is a block diagram illustrating a configuration for determining a prediction mode by the SAD method.

  The prediction value calculation units 11 to 13 calculate the prediction value of the prediction target image of the prediction unit with reference to the reference image stored in the frame buffer 10 in a predetermined prediction mode. The reference image is a decoded image (reconstructed image) obtained by locally decoding an encoded frame image. The subtractors 14 to 16 calculate a prediction error from the prediction target image (input image) of the prediction unit and the prediction value, and store the prediction error in the memories 17 to 19.

  The SAD calculation units 20 to 22 calculate the prediction error SAD stored in the memories 17 to 19 as an evaluation value. The processing so far is performed in parallel for the number of prediction modes, and SADs for the number of prediction modes are calculated.

  The selection unit 23 selects a prediction mode corresponding to the SAD indicating the minimum value among the calculated SADs. The prediction error selection unit 24 reads out the prediction error corresponding to the prediction mode selected by the selection unit 23 from the memory 17, 18 or 19, and outputs it.

  That is, in the SAD scheme, SADs corresponding to the number of prediction modes are calculated, and the prediction mode is determined based on the SAD. As described above, there are nine types of intra 4 × 4 prediction modes and four types of intra 16 × 16 prediction modes. In other words, when predicting luminance information, there are 13 types of prediction modes. FIG. 1 shows a case where there are three types of prediction modes for the sake of simplicity. Hereinafter, the case where there are three types of prediction modes will be described.

  FIG. 2 is a block diagram illustrating a configuration for determining a prediction mode by the RDO method. Note that components substantially the same as those in the SAD scheme shown in FIG. 1 are denoted by the same reference numerals, and detailed description thereof is omitted.

The evaluation value calculation units 25 to 27 calculate the evaluation value J using Expression (1) based on the prediction errors stored in the memories 17 to 19.
J = D + λ (QP) × R (1)
Where D is the coding distortion,
R is the code length of the prediction target image,
QP is the quantization scale,
λ is a Lagrange multiplier with QP as a parameter.

  In Equation (1), the encoding distortion D is calculated as the sum of squares of the difference between the decoded image that has been locally decoded after encoding the prediction target image and the prediction target image. As the quantization scale QP increases (at a high bit rate), the value of the Lagrange multiplier λ also increases with a power.

  Similar to the SAD method, the selection unit 23 selects a prediction mode corresponding to the evaluation value indicating the minimum value among the calculated evaluation values. The prediction error selection unit 24 reads out the prediction error corresponding to the prediction mode selected by the selection unit 23 from the memory 17, 18 or 19, and outputs it.

  Comparing the RDO scheme and the SAD scheme, the RDO scheme that selects the optimal prediction mode from the code length R and the coding distortion D shows higher coding efficiency. However, as shown in Expression (1), the amount of calculation of the coding distortion D and the code length R is very large, and the implementation cost is very high. On the other hand, the SAD method is easy to implement, but the coding efficiency tends to deteriorate by 5% or more compared to the RDO method because the code amount is not considered.

JP 2003-299102 JP

  An object of the present invention is to provide an encoding that is easy to implement and enables efficient intra-frame prediction.

  The present invention has the following configuration as one means for achieving the above object.

  The present invention is an encoding method of a moving image that performs intra-frame prediction, wherein the prediction for each prediction mode is performed with reference to the encoded decoded image of the prediction unit adjacent to the prediction target image of the prediction unit. Calculate the prediction error of the target image, calculate the absolute value sum of the prediction errors of the prediction target image for each of the prediction modes, and based on the prediction mode information used for encoding the decoded image adjacent to the prediction target image The syntax length of the prediction target image is calculated, and the prediction mode is selected for the prediction encoding of the prediction target image based on the prediction mode information used for encoding the decoded image adjacent to the prediction target image. Probability is calculated, and each of the prediction modes is calculated from the absolute value sum, the syntax length, the probability, and the quantization scale corresponding to the prediction target image. Calculating an evaluation value based on the evaluation value of the prediction mode, respectively, and selects the prediction mode, based on the result of the selection, and outputs the prediction error selectively.

  According to the present invention, it is possible to obtain coding that is easy to implement and enables efficient intra-frame prediction.

  Hereinafter, the moving image processing of the embodiment according to the present invention will be described in detail with reference to the drawings.

[Device configuration]
FIG. 3 is a block diagram illustrating a basic configuration example of the moving image encoding apparatus according to the embodiment. 1 and 2 are denoted by the same reference numerals, and detailed description thereof is omitted.

  The intra prediction unit 31 refers to the local decoded image stored in the frame buffer 10, performs intra-frame prediction of the prediction target image for each prediction unit, and outputs a prediction error of the prediction target image. The inter prediction unit 32 refers to the locally decoded frame image 33, performs inter-frame prediction of the prediction target image for each frame unit, and outputs a prediction error of the prediction target image. The frame image 33 to be referenced is stored in the frame buffer.

  The integer conversion unit 34 receives the prediction error, performs integer conversion, and outputs a conversion coefficient. The quantization unit (Q) 35 quantizes the transform coefficient. The entropy encoding unit 36 entropy encodes the quantized transform coefficient. The buffer 37 buffers the code output from the entropy encoding unit 36 and outputs it as a code stream outside the apparatus.

The inverse quantization unit (Q ⁻¹ ) 38 inversely quantizes the quantized transform coefficient. The inverse integer transform unit 39 performs inverse integer transform on the inversely quantized transform coefficient. The reconstructed pixel calculation unit 40 calculates the decoded pixel by adding the prediction value of the intra prediction unit 31 or the prediction value of the inter prediction unit 32 and the prediction error restored by inverse integer transformation. The decoded pixel is stored in the frame buffer 10 and used as a reference image. The deblocking filter 41 is a filter that reduces block noise.

[Intra prediction section]
FIG. 4 is a block diagram illustrating a configuration example of a part of the intra prediction unit 31. In addition, the same code | symbol is attached | subjected to the structure substantially the same as FIGS. 1-3, and the detailed description is abbreviate | omitted. As will be described later, the intra prediction unit 31 includes the configuration illustrated in FIG. 4 for each intra 4 × 4, intra 16 × 16, and intra chroma.

  The absolute value sum calculation units (SAD calculation units) 107a to 107c calculate SADs of prediction errors stored in the memories 17 to 19. In the present embodiment, the processing may be performed without using the memories 17 to 19, and in this case, the memories 17 to 19 are unnecessary.

  The syntax length calculation units 108a to 108c calculate the syntax length corresponding to the prediction mode number from the prediction mode information stored in the reference prediction mode memory 102 and applied to the encoded image block adjacent to the prediction target image. calculate.

  The probability calculation units 109a to 109c calculate the probability that the prediction mode is selected from the prediction mode information applied to the encoded image block adjacent to the prediction target image stored in the reference prediction mode memory 102. .

The evaluation value calculation units 110a to 110c calculate the evaluation value of the prediction value calculation formula using the following formula.
Cost = [{λ (QP) × SAD + 1} ≪ Rh] + Cost _th … (2)
Here, SAD is SAD input from the corresponding SAD calculation unit,
Rh is the syntax length input from the corresponding syntax length calculation unit,
Cost _th is an offset according to the selection probability input from the corresponding probability calculation unit,
λ (QP) is a Lagrange multiplier with the quantization scale QP as a parameter.

In Expression (2), “<<” represents a bit shift, and “y << x” means that y is shifted left by x bits, in other words, y is multiplied by 2 ^× .

  The selection unit 111 selects a prediction mode corresponding to the evaluation value indicating the minimum value among the calculated evaluation values, and stores information indicating the selection result (number of the prediction value arithmetic expression) in the reference prediction mode memory 102. To do. Also, the sum of the evaluation values calculated by the evaluation value calculation units 110a to 110c is output. The prediction error selection unit 24 reads the prediction error corresponding to the prediction mode selected by the selection unit 111 from the memory 17, 18 or 19, and selectively outputs it.

  FIG. 5 is a block diagram illustrating a configuration example of the intra prediction unit 31.

  The selection unit 124 inputs the evaluation value sum output from the intra 4 × 4 prediction unit 121 for 16 image blocks, and the evaluation value sum output from the intra 16 × 16 prediction unit 122 as one image block (MB). Enter minutes. Then, the sum ΣA (sum) of the evaluation value sums for 16 image blocks is compared with the evaluation value sum A for 1 MB. Then, if ΣA> A, the prediction error selection unit 125 is selected so that the prediction error output by the intra 16 × 16 prediction unit 122 is selected, and if ΣA ≦ A, the prediction error output by the intra 4 × 4 prediction unit 121 is selected. Control. That is, a prediction unit for luminance information is selected in units of 16 × 16 pixel blocks.

  Then, the prediction error selection unit 125 outputs a prediction error of luminance information. Further, the intra chroma prediction unit 123 outputs a prediction error of chroma information.

  The moving image processing according to the second embodiment of the present invention will be described below. Note that the same reference numerals in the second embodiment denote the same parts as in the first embodiment, and a detailed description thereof will be omitted.

  FIG. 6 is a block diagram illustrating a partial configuration example of the intra prediction unit 31 according to the second embodiment. In addition, the same code | symbol is attached | subjected to the structure substantially the same as FIGS. 1-4, and the detailed description is abbreviate | omitted. Further, the intra prediction unit 31 has the configuration shown in FIG. 6 for each intra 4 × 4, intra 16 × 16, and intra chroma.

Based on the SAD calculated by the corresponding SAD calculation units 107a to 107c, the code length calculation units 112a to 112c derive the code length R _ecs [SAD] from the code length table using the SAD as an index.

The evaluation value calculation units 110a to 110c calculate the evaluation value of the prediction value calculation formula using the following formula.
Cost = λ (QP) × R _ecs [SAD] + Rh + Cost _th … (3)
Here, R _ecs [SAD] is the code length input from the corresponding code length calculation unit,
Rh is the syntax length input from the corresponding syntax length calculation unit,
Cost _th is an offset according to the selection probability input from the corresponding probability calculation unit,
λ (QP) is a Lagrange multiplier with the quantization scale QP as a parameter.

  Note that the overall configuration of the intra prediction unit 31 is the same as the configuration of the first embodiment illustrated in FIG.

  According to the configuration of the intra prediction unit 31 shown in the first and second embodiments, in intra prediction compliant with H.264, intra 4 × 4 is the same prediction value calculation formula (prediction direction) as the adjacent encoded image block. ) To reduce the code amount. Also, when the quantization scale QP is large, the intra 16 × 16 selection rate can be increased to reduce the code amount without degrading the image quality. In other words, it is possible to capture intra-image characteristics and perform intra prediction encoding with high encoding efficiency.

  In this embodiment, in order to simplify the description, the intra 4 × 4 and the intra 16 × 16 are used for the luminance intra prediction. However, in addition to the intra 4 × 4 and the intra 16 × 16, the intra 8 × 8 is used. May be used. The prediction unit is not limited to 4 × 4 pixels, 8 × 8 pixels, and 16 × 16 pixels, as long as the size can divide the macroblock equally.

[Other embodiments]
Note that the present invention can be applied to a system constituted by a plurality of devices (for example, a computer, an interface device, a reader, a printer, etc.), but an apparatus (for example, a copier, a facsimile machine, a control device) composed of a single device. Etc.).

  Another object of the present invention is to supply a recording medium or a recording medium recording a computer program for realizing the functions of the above embodiments to a system or apparatus. This can also be achieved by the computer (CPU or MPU) of the system or apparatus executing the computer program. In this case, the software read from the recording medium itself realizes the functions of the above embodiments, and the computer program and the computer-readable recording medium storing the computer program constitute the present invention. .

  Further, the above functions are not only realized by the execution of the computer program. That is, according to the instruction of the computer program, the operating system (OS) and / or the first, second, third,... This includes the case where the above function is realized.

  The computer program may be written in a memory of a device such as a function expansion card or unit connected to the computer. That is, it includes the case where the CPU of the first, second, third,... Device performs part or all of the actual processing according to the instructions of the computer program, thereby realizing the above functions.

  When the present invention is applied to the recording medium, the recording medium stores a computer program corresponding to or related to the flowchart described above.

A block diagram for explaining a configuration for determining a prediction mode by the SAD method, Block diagram for explaining the configuration for determining the prediction mode by the RDO method, A block diagram showing a basic configuration example of a moving image encoding device of an embodiment, A block diagram showing a configuration example of a part of the intra prediction unit, A block diagram showing a configuration example of an intra prediction unit, Block diagram showing a configuration example of a part of the intra prediction unit of the second embodiment, It is a figure which shows the predicted value calculation type | formula syntax when intra 16x16 is selected.

Claims (8)

A video encoding device that performs intra-frame prediction,
A calculation unit that calculates a prediction error of the prediction target image for each prediction mode with reference to the encoded decoded image of the prediction unit adjacent to the prediction target image of the prediction unit;
Absolute value sum calculating means for calculating an absolute value sum of prediction errors of the prediction target image for each of the prediction modes;
Syntax length calculation means for calculating a syntax length of the prediction target image based on information on a prediction mode used for encoding a decoded image adjacent to the prediction target image;
Probability calculating means for calculating a probability that the prediction mode is selected for predictive encoding of the prediction target image based on information on a prediction mode used for encoding a decoded image adjacent to the prediction target image;
Evaluation value calculating means for calculating an evaluation value for each of the prediction modes from the absolute value sum, the syntax length, the probability, and a quantization scale corresponding to the prediction target image;
Selection means for selecting the prediction mode based on the evaluation value for each of the prediction modes;
And an output unit that selectively outputs the prediction error based on the selection result.   2. The encoding apparatus according to claim 1, wherein the selection unit selects the prediction mode corresponding to a minimum evaluation value. Furthermore, from the absolute value sum for each of the prediction modes, the code length calculation means for calculating the code length for each of the prediction modes,
3. The encoding apparatus according to claim 1, wherein the evaluation value calculation unit uses the code length instead of the absolute value sum.   4. The encoding apparatus according to claim 3, wherein the code length is calculated from the sum of absolute values with reference to a table. And calculating means for calculating the sum of the evaluation values for each size of the prediction unit;
4. The method according to claim 1, further comprising a determination unit that compares the sum of evaluation values calculated for each prediction unit by the calculation unit and determines the prediction unit of the prediction target image. The encoding device according to one item. A video encoding method for performing intra-frame prediction,
With reference to the encoded decoded image of the prediction unit adjacent to the prediction target image of the prediction unit, the prediction error of the prediction target image for each prediction mode is calculated,
Calculating an absolute value sum of prediction errors of the prediction target image for each of the prediction modes;
Based on the prediction mode information used for encoding the decoded image adjacent to the prediction target image, the syntax length of the prediction target image is calculated,
Based on the prediction mode information used for encoding the decoded image adjacent to the prediction target image, calculating the probability that the prediction mode is selected for prediction encoding of the prediction target image,
From the absolute value sum, the syntax length, the probability, and a quantization scale corresponding to the prediction target image, an evaluation value for each of the prediction modes is calculated,
Based on the evaluation value for each of the prediction modes, select the prediction mode,
An encoding method, wherein the prediction error is selectively output based on the selection result.   6. A non-transitory computer-readable storage medium storing a program for controlling a computer device to function as each unit of the encoding device according to claim 1.   8. A computer-readable recording medium on which the program according to claim 7 is recorded.

Images