KR20080013843A - Method and apparatus for loseless encoding and decoding image - Google Patents

Abstract

A lossless encoding and decoding method of a moving picture and an apparatus thereof are provided to calculate the difference between pixels of the inside of a residual block by the inter prediction result and to code formed differential residual blocks. A prediction about an input image is performed with a predetermined block size and then a prediction block corresponding to the present block for coding is produced. A residual block consisting of residual signals corresponding to the difference between the pixels of the prediction block and the pixels corresponding to the present block is produced(720). The difference value between the residual signals for constructing the residual block is calculated according to the predetermined direction and the differential residual block is produced on the basis of the calculated difference value. Thereafter, the entropy coding of the differential residual block is performed(730).

Description

Method and apparatus for lossless encoding and decoding of video {Method and apparatus for loseless encoding and decoding image}

The present invention relates to encoding and decoding of video data, and more particularly, to a lossless video encoding having a higher compression ratio by encoding a differential residual block formed by calculating a difference between pixels in a residual block based on an inter prediction result. The present invention relates to a decoding method and an apparatus thereof.

H.264 / MPEG-4 Advanced Video Coding (AVC), one of the compression standards for video signals, provides multiple reference motion compensation, loop filtering, and variable block size motion compensation. Various techniques are adopted to increase the compression efficiency, such as compensation) and entropy coding such as CABAC.

According to the H.264 standard, encoding and decoding are performed in units of a plurality of macroblocks included in one frame or subblocks obtained by dividing or dividing a macroblock. Each macroblock is encoded in all encoding modes available for inter prediction and intra prediction, and then the bit rate required for encoding the block and the degree of distortion of the original block and the decoded block are encoded. Therefore, a block is encoded by selecting an optimal encoding mode. Here, inter prediction refers to performing prediction by referring to a macro block of an adjacent frame in predicting a motion of a macro block of a current frame, and intra prediction refers to a macro block of a current frame to be encoded to be adjacent to a macro block in the frame. It is to perform the prediction by using.

On the other hand, there are lossy coding and lossless coding in video encoding. In order to losslessly encode a video, an error signal that is a difference between a prediction block and a current block according to a prediction result is entropy coded without DCT (Discrete Cosine Transform) or quantization. According to the related art, the compression rate is performed by performing entropy encoding on the residual block, which is the error signal, without performing a separate process, or entropy encoding the difference value formed by using pixel values of neighboring blocks adjacent to the current residual block. This falls, and there is a difficulty in parallel processing using the pipeline technique.

Accordingly, the present invention has been made to solve the above problems, and generates a differential residual block by calculating a difference between pixels in the residual block that is output as a result of performing inter prediction in units of blocks having a predetermined size. It is an object of the present invention to provide a lossless video encoding and decoding method and apparatus for increasing a compression rate by encoding the differential residual block.

In order to solve the above technical problem, a method of lossless encoding of a video according to the present invention comprises: generating a prediction block corresponding to a current block to be encoded by performing inter prediction on a unit of a block having a predetermined size between a reference picture and a current picture; ; Generating a residual block consisting of a residual signal corresponding to a difference between pixels of the prediction block and corresponding pixels of the current block; Calculating a difference value between residual signals constituting the residual block according to a predetermined direction, and generating a differential residual block based on the calculated difference value; And entropy encoding the differential residual block.

A lossless encoding apparatus of a video according to the present invention includes an inter prediction unit configured to generate a prediction block corresponding to a current block to be encoded by performing inter prediction between a reference image and a current image in units of blocks having a predetermined size; A subtraction unit configured to generate a residual block including a residual signal corresponding to a difference between pixels of the prediction block and corresponding pixels of the current block; A difference residual conversion unit calculating a difference value between residual signals constituting the residual block according to a predetermined direction, and generating a difference residual block based on the calculated difference value; And an entropy encoding unit for entropy encoding the differential residual block.

According to the present invention, a lossless decoding method of a video may include performing an entropy coded bitstream based on a differential residual block generated by calculating a difference between pixels in a residual block that is a difference between an inter predicted prediction block and a current block. Receiving; Entropy decoding the bitstream to decode a differential residual block; Generating the residual block by performing inverse transform on the decoded differential residual block; And reconstructing the image by adding an inter predicted block by using the residual block and the image data included in the bitstream.

The lossless decoding apparatus of a video according to the present invention may perform an entropy coded bitstream based on a differential residual block generated by calculating a difference between pixels in a residual block that is a difference between an inter predicted prediction block and a current block. An entropy decoding unit for receiving and entropy decoding the bitstream to output a differential residual block; A differential residual transform unit generating the residual block by performing an inverse transform on the differential residual block; An inter prediction unit generating an inter predicted block by using the image data included in the bitstream; And an adder configured to reconstruct an image by adding the residual block and the inter predicted block.

According to the present invention, there is provided a method for lossless encoding of a video, comprising: generating a prediction block corresponding to a current block to be encoded by performing prediction on a unit of a predetermined size with respect to an input image; Generating a residual block consisting of a residual signal corresponding to a difference between pixels of the prediction block and corresponding pixels of the current block; Calculating a difference value between residual signals constituting the residual block according to a predetermined direction, and generating a differential residual block based on the calculated difference value; And entropy encoding the differential residual block.

The lossless decoding method of a video according to the present invention comprises a residual block composed of a residual signal corresponding to a difference between a corresponding pixel of a prediction block and a current block in a predetermined direction, and between residual signals of the residual block. Calculating a difference value of and receiving an entropy-encoded bitstream using a differential residual block generated based on the calculated difference value; Entropy decoding the bitstream to decode the differential residual block; Generating the residual block by performing inverse transform on the decoded differential residual block; And reconstructing the image by adding a predicted block by using the residual block and the image data included in the bitstream.

According to the present invention as described above, it is possible to improve the compression rate during lossless coding of the residual block output as a result of inter prediction. In addition, according to the present invention, by forming a differential residual block using pixels in the residual block to perform entropy encoding, unlike the conventional technology using the pixel of the previous block, the processing speed is increased through parallel processing using the pipeline technique. The compression ratio can be improved without any change.

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

1 is a block diagram showing the configuration of a video encoding apparatus according to the present invention.

Referring to FIG. 1, the image encoding apparatus 100 according to the present invention may include a motion estimator 102, a motion compensator 104, an intra predictor 106, a subtractor 107, and a differential residual transform unit ( 108, transform / quantization unit 110, entropy encoding unit 112, differential residual inverse transform unit 114, inverse transform / inverse quantization unit 116, adder 117, frame memory 118, and filter ( 120).

The motion estimator 102 performs a motion estimation to read an area most similar to a current macro block in an image of a reference frame stored in the frame memory 118. Specifically, referring to FIG. 2 illustrating a process of performing motion estimation by the motion estimator 102 according to the present invention, the motion estimator 102 includes a predetermined region centered on the position of the current macroblock in a reference frame. Is searched to select the area in which the difference with the current macroblock is the smallest in the search area as the most similar area. In addition, the motion estimator 102 outputs a spatial position difference between the most similar area and the current macro block in the form of a motion vector.

The motion compensator 104 generates an inter predicted prediction block by reading an area most similar to a current macro block from an image of a reference frame stored in the frame memory 118 using the motion vector. As such, the motion estimator 102 and the motion compensator 104 operate as an inter predictor for performing inter prediction.

The intra prediction unit 106 performs intra prediction to find a prediction value of the current block in the current picture.

The inter prediction and intra prediction processes may be performed on not only 16 × 16 macro blocks but also 16 × 8, 8 × 16, 8 × 8, 8 × 4, 4 × 8, and 4 × 4 blocks. .

When inter prediction or intra prediction is performed to form a prediction block corresponding to a block to be currently encoded, the subtractor 107 calculates a difference between the current block and the prediction block to form a residual block Dn. Output In general, the residual is encoded in order to reduce the amount of data during encoding of the image.

In the case of lossy coding, the residual block undergoes discrete cosine transformation and quantization in the transform / quantization unit 110, and the quantized residual block is entropy encoded in the entropy encoder 112 and output as a bitstream.

In the case of lossless encoding, the residual block is entropy encoded without passing through the transform / quantization unit 110. In particular, the image encoding apparatus 100 according to the present invention converts the residual block output as an inter prediction result into a differential residual dual block through a differential residual transform unit 108, and converts the differential residual block into the entropy encoding. The unit 112 performs entropy encoding and outputs the bitstream.

3 is a diagram illustrating a residual block transformed by the differential residual converter 108 according to the present invention.

As described above, the prediction block inter predicted by the motion estimator 102 and the motion compensator 104 is subtracted from the current block, and as a result, the residual block 300 is output. In FIG. 3, an inter predicted residual block corresponding to a 16 × 16 macroblock is shown, and pixel values located in the x-th row and the y-th column of the residual block are denoted as r _{x , y} .

The difference residual converter 108 receives a residual block having a predetermined size as shown in FIG. 3 and calculates a difference between pixels constituting the residual block 300 in a predetermined direction. Create a differential residual block. That is, a differential residual block is generated by performing differential pulse code modulation (DPCM) in a horizontal or vertical direction between the pixels constituting the residual block 300. The reason for generating the differential residual block as described above is to improve the compression ratio by reducing the size of the pixel value constituting the residual block 300.

4 is a view illustrating a process of generating a differential residual block in a horizontal direction in the differential residual transform unit 106 according to the present invention, and FIG. 5 is a vertical direction in the differential residual transform unit 106 according to the present invention. FIG. 6 is a diagram illustrating a process of generating a differential residual block of FIG. 6, and FIG. 6 is a diagram illustrating a differential residual block generated by the difference residual converter 106.

Referring to FIG. 4, the difference residual converter 108 includes horizontal pixels formed by subtracting neighboring pixel values adjacent to a horizontal direction with respect to each pixel of the residual block 300 of FIG. 3. The differential residual block 400 is formed. That is, the pixel values in the x th line, the y-th column in the residual block r _x, speaking _y, the pixel value of the horizontal difference residual block (400) x-th line, the y-th column in r _{'x, y} Is calculated as in Equation 1 below.

Here, the pixels located in the first column of the difference residual block 400, that is, r ' _0,0 , r' _{1 , 0} , r ' _{2 , 0} , ..., r' _{15 , 0} are each The values of the pixels located in the first column of the residual block 300 corresponding to the pixels are maintained as they are. In other words, r ' _{N, 0} = r _{N, 0} (N = 0,1,2, ..., 15).

In addition, referring to FIG. 5, the difference residual transform unit 108 is composed of pixels generated by subtracting neighboring pixel values adjacent to the vertical direction with respect to each pixel of the residual block 300 of FIG. 3. To form a vertical differential residual block 500. That is, the pixel values r ' _{x , y} located in the x-th row and the y-th column of the vertical residual residual block 500 are calculated as in Equation 2 below.

Here, the pixels located in the first row of the difference residual block 400, that is, r ' _0,0 , r' _{0 , 1} , r ' _{0 , 2} ,..., R' _{0 , 15} are described above. The values of the pixels located in the first row of the residual block 300 corresponding to each pixel are maintained as they are. In other words, r ' _{0, N} = r _{0, N} (N = 0, 1, 2, ..., 15).

* The horizontal or vertical differential residual block 400 or 500 generated by the differential residual transform unit 108 is directly input to the entropy encoder 112 without undergoing a transformation and quantization process and is lossless using CABAC or the like. Is encoded.

On the other hand, the difference residual block output from the difference residual conversion unit 108 is inversely converted into a residual block as shown in FIG. 3 by the difference residual inverse conversion unit 114, and is added by the adder 117. The prediction block is added and reconstructed. Here, the inverse transform on the difference residual block may be performed by summing pixels located in a predetermined direction with respect to pixels to be inversely transformed in the difference residual block. Specifically, the entropy-encoded differential residual block is adjacent to each other in the horizontal direction. Is generated from the difference between the inverse of the difference residual block of the x-th row between, y-th pixel value in the column r _{'x, y} when the LA, the difference residue inverse transform unit 114 x of the residual block The pixel values r _{x and y} located in the first row and the y th column are calculated by the following equation.

In addition, pixels in which the entropy-encoded differential residual block is adjacent in the vertical direction When the difference is generated from the difference, the difference residual inverse transform unit 114 calculates the pixel values r _{x, y} located in the x-th row and the y-th column of the residual block through Equation 4 below. .

On the other hand, the residual block quantized according to the conventional encoding method, not according to the encoding method according to the present invention, is recovered through the inverse transform / inverse quantization unit 116. The reconstructed block passes through a filter 120 that performs deblocking filtering, and is then stored in the frame memory 118 and used to perform inter prediction on the next picture.

7 is a flowchart illustrating a lossless encoding method of a video according to the present invention.

Referring to FIG. 7, in operation 710, an inter predictor including a motion estimator 102 and a motion compensator 104 performs inter prediction between a reference image and a current image in units of blocks of a predetermined size and encodes the current block. A predictive block corresponding to the PB is formed and a residual block corresponding to the difference between the predicted block and the current block is output.

In operation 720, a difference residual block is generated by calculating a difference between the pixels constituting the residual block according to a predetermined direction. Equations 1 and 2 described above may be used to generate the differential residual block.

In step 730, the bitstream is output by performing entropy encoding on the differential residual block.

8 is a diagram schematically illustrating a format of a bitstream output from an encoding apparatus according to the present invention.

Referring to FIG. 8, the bitstream 800 is divided into a header 810 including encoding mode information and the like, and a data portion 820 including information about transform coefficients and motion vectors. When entropy encoding is performed using the differential residual block, performance may be worse than that of entropy encoding using a residual block which has not been converted conventionally in terms of compression rate. Accordingly, the entropy encoder 112 compares the cost of the mode of entropy encoding the differential residual block with the cost of the mode of entropy encoding by transforming and quantizing the residual block, and finalizes a mode having a smaller cost. Decide on the normal coding mode. In addition, by adding a predetermined bit 810a indicating the final encoding mode to the header of the bitstream, it is known how the bitstream currently received by the decoding apparatus that has received the bitstream is encoded.

In addition, when the entropy encoding unit 112 performs entropy encoding on the difference residual block, the entropy encoder 112 adds information indicating a predetermined direction in which the adjacent pixels are subtracted when generating the difference residual block to a header of the bitstream. can do. For example, a bit 810b indicating whether the difference residual block is formed by subtracting pixels in the residual block in a horizontal or vertical direction may be added to the header of the bitstream.

9 is a block diagram showing the configuration of a video lossless decoding apparatus according to the present invention.

Referring to FIG. 9, the decoding apparatus 900 according to the present invention includes an entropy decoder 910, a differential residual inverse transform unit 920, an inverse quantization / inverse transform unit 930, an adder 940, and an intra predictor. 950 and a motion compensator 960.

When the entropy decoding unit 910 receives the bitstream encoded according to the lossless encoding method according to the present invention, the entropy decoding unit 910 entropy decodes the bitstream and outputs a differential residual block. In addition, when receiving the bitstream encoded according to the prior art, the entropy decoding unit 910 entropy decodes the bitstream and outputs a residual block. As described above, the entropy decoding unit 910 may determine how the current bitstream to be encoded is encoded from the encoding information included in the header of the bitstream.

The difference residual inverse transform unit 920 generates a residual block by performing an inverse transform on the input difference residual block. The difference residual inverse transform unit 920 operates in the same manner as the difference residual inverse transform unit 114 of FIG. 1. Specifically, when the differential residual block 600 decoded by the entropy decoding unit 910 as shown in FIG. 6 is inputted, in the predetermined direction subtracted during generation of the pixels of the differential residual block 600. Accordingly, a residual block is generated using 3 or 4 in the math. Here, when the difference residual block 600 is generated through subtraction between pixels adjacent to each other in the horizontal direction, Equation 3 is used, and when the difference residual block 600 is generated through subtraction between pixels adjacent to each other in the vertical direction, Equation 4 may be used. For example, when the difference residual block 600 as shown in FIG. 6 is generated through subtraction between adjacent pixels in the horizontal direction, each pixel value r of the residual block from the difference residual block 600 is generated. _{x, y} ) can be calculated as

r _{N , 0} = r ' _{N , 0} (N = 0,1,2, ..., 15)

r _{N , 1} = r ' _{N , 0} + r' _{N , 1}

r _{N , 2} = r ' _{N , 0} + r' _{N , 1} + r ' _{N , 2}

.......

r _{N , 15} = r ' _{N , 0} + r' _{N , 1} + r ' _{N , 2} + r' _{N , 3} + ... + r ' _{N , 11} + r' _{N , 12} + r ' _{N , 13} + r ' _{N , 14} + r' _{N , 15}

Similarly, when the difference residual block 600 as shown in FIG. 6 is generated through subtraction between adjacent pixels in the vertical direction, each pixel value r _x of the residual block from the difference residual block 600 is generated. _{, y} ) can be calculated as

r _{0 , N} = r ' _{0 , N} (N = 0,1,2, ..., 15)

r _{1 , N} = r ' _{0 , N} + r' _{1 , N}

r _{2 , N} = r ' _{0 , N} + r' _{1 , N} + r ' _{2 , N}

.......

r _{15 , N} = r ' _{0 , N} + r' _{1 , N} + r ' _{2 , N} + r' _{3 , N} + ... + r _{'11 , N} + r' _{12 , N} + r _{'13 , N} + r _{'14 , N} + r' _{15 , N}

The motion compensator 960 generates an inter predicted block by using image data included in the bitstream. The inter predicted block is added to the adder 940 with the reconstructed residual block to reconstruct an image.

Meanwhile, the inverse quantization / inverse transformer 930 generates a residual block by performing inverse quantization and inverse transformation on the bitstream encoded by the lossy coding method, and the intra predictor 950 and the motion compensator 960. ) Generates a prediction block according to the encoded picture type using the decoded header information, and the generated prediction block is added to the residual block through the adder 940 to reconstruct an image.

10 is a flowchart illustrating a lossless decoding method of a video according to the present invention.

Referring to FIG. 10, in step 1010, the entropy decoding unit 910 receives a bitstream encoded according to a lossless encoding method according to the present invention, entropy decodes the bitstream, and outputs a differential residual block.

In operation 1020, the differential residual inverse transform unit 920 generates a residual block by performing inverse transform on the decoded differential residual block. As described above, Equation 3 or 4 may be used in consideration of a predetermined direction in which pixels are subtracted when generating the differential residual block.

In operation 1030, the motion compensator 960 generates a prediction block by performing inter prediction using image data included in the bitstream.

In operation 1040, the adder 940 reconstructs an image by adding the residual block generated by the difference residual inverse transform unit 920 and the prediction block generated by the motion compensator 960.

Table 1 is a table comparing the compression rate when the video is compressed by the encoding method according to the prior art and the encoding method according to the present invention for the JVT test sequences configured in the RGB 4: 4: 4 format. In experiments, horizontal or vertical differential residual blocks were entropy coded in units of 16 × 16 macroblocks.

Sequences Total bits Saving bits Saving Ratio Prior art Invention analog TV 2869478616 2455621474 413857142 14.42% cardtoss 2588812120 2374566808 214245312 8.28% dinner 2582247432 2412085245 170162187 6.59% bicycle 1864307320 1760634247 103673073 5.56% tomatoes 2614905896 2276470749 338435147 12.94% Average - - 248074572 9.56%

Referring to Table 1, it can be seen that when encoding the differential residual block according to the present invention compared to the encoding method according to the prior art, the average compression ratio is increased by about 9.5%.

On the other hand, the above-described video encoding and decoding method can be created by a computer program. Codes and code segments constituting the program can be easily inferred by a computer programmer in the art. In addition, the program is stored in a computer readable media, and read and executed by a computer to implement a video encoding and decoding method. The information storage medium includes a magnetic recording medium, an optical recording medium, and a carrier wave medium.

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

1 is a block diagram showing the configuration of a video encoding apparatus according to the present invention.

FIG. 2 is a diagram illustrating a process of performing motion estimation by the motion estimation unit of FIG. 1.

FIG. 3 is a diagram illustrating a residual block input to the differential residual converter of FIG. 1.

FIG. 4 is a diagram illustrating a process of generating a differential residual block in a horizontal direction in the difference residual converter of FIG. 1.

FIG. 5 is a diagram illustrating a process of generating a differential residual block in a vertical direction in the difference residual converter of FIG. 1.

FIG. 6 is a diagram illustrating a differential residual block generated by the differential residual converter of FIG. 1.

7 is a flowchart illustrating a lossless encoding method of a video according to the present invention.

8 is a diagram schematically illustrating a format of a bitstream output from an encoding apparatus according to the present invention.

9 is a block diagram showing the configuration of a video lossless decoding apparatus according to the present invention.

10 is a flowchart illustrating a lossless decoding method of a video according to the present invention.

Claims (2)

In the lossless encoding method of a video, Generating a prediction block corresponding to the current block to be encoded by performing prediction on the input image in units of blocks having a predetermined size; Generating a residual block consisting of a residual signal corresponding to a difference between pixels of the prediction block and corresponding pixels of the current block; Calculating a difference value between residual signals constituting the residual block according to a predetermined direction, and generating a differential residual block based on the calculated difference value; And And entropy encoding the differential residual block. In the lossless decoding method of a video, Comprising a residual block consisting of a residual signal corresponding to the difference between the prediction block and the corresponding pixels of the current block according to a predetermined direction, calculates the difference value between the residual signal of the residual block, and calculated Receiving an entropy coded bitstream using a differential residual block generated based on the difference value; Entropy decoding the bitstream to decode the differential residual block; Generating the residual block by performing inverse transform on the decoded differential residual block; And And reconstructing the image by adding the predicted block by using the residual block and the image data included in the bitstream.

Images