KR0152017B1 - Coding and decoding system for adaptable process - Google Patents

Abstract

Means for encoding image data divided into blocks of a predetermined size, and means for decoding the encoded data for performing predictive encoding, the encoding system for outputting the encoded data and the image encoded by the encoding system. In a decoding system for decoding data, the image data output from the decoding means in the encoding system and the image data decoded in the decoding system have a similar degree of damage to the image data due to additional noise such as quantization noise. The coefficient data capable of restoring the data is adaptively generated and transmitted to the decoding system, and the decoding system uses the coefficient data to restore the damaged image data to be similar to the previous image data.

Description

Coding and Decoding System Using Adaptive Processing

1 is a block diagram showing an example of a conventional video data encoding apparatus.

2 is a block diagram illustrating an example of a conventional image data decoding apparatus.

3 is a block diagram showing an embodiment of an image data encoding apparatus according to the present invention.

4 is a conceptual diagram showing a correlation between a current pixel and an adjacent pixel.

5 is a block diagram showing an embodiment of an image data decoding apparatus according to the present invention.

* Explanation of symbols for main parts of the drawings

11 DCT unit 12 quantization unit

13,20 variable variable encoding unit 14 buffer

15,25: reverse quantization unit 16,26: reverse DCT unit

17,28: frame memory 18: motion estimation

19,29: motion compensation 24,27: variable length decoding

21: data combiner 22: data separator

31: delay unit 32: adaptive processing unit

33: post-processing unit

The present invention relates to a system for encoding and decoding video signals, and more particularly, to an encoding and decoding system for providing more improved image quality by adaptively processing a corrupted signal including noise in encoding and decoding.

In general, as a method of compressing data to store or transmit image data, it is used in a DCT (Discrete Cosine Transform) encoding method using motion compensation.

In addition to DCT, there are various methods for encoding video signals. Recently, ICs having processing speeds capable of processing video signals have been developed due to the development of digital signal processing technology. These coding methods are currently being standardized by organizations such as CCITT under ISO and CCIR in Europe.

FIG. 1 shows a predictive coding system using motion compensation as an example of a conventional coding apparatus. The encoding apparatus of FIG. 1 includes means for quantizing a transform coefficient after performing DCT transform on input image data, means for further compressing the data amount by variable-length encoding the quantized data, and keeping the input / output data amount constant. Inverse quantization and inverse transformation of the quantized data, a buffer for outputting a predetermined quantization stepize signal for controlling the quantization means, and motion estimation and motion compensation are performed from the inverse transformed reconstructed picture and the block to be encoded. It consists of a means.

The image data of the N × N block is input through the input terminal 10, and the block image data and the image data fed back calculate the error data in the first adder A1. The error data is converted into data in the frequency domain by the DCT unit 11, and the energy of the converted conversion coefficient is mainly collected toward the low frequency side. The quantization unit 12 converts the transform coefficients into representative values of a predetermined level through a predetermined quantization process. Then, the variable length encoding unit 13 compresses the data in a variable length encoding manner by utilizing the statistical characteristics of the representative values and outputs the compressed image data (V _CD ). The output data of the first variable length coding unit 13 is supplied to the buffer 14, and the buffer 14 outputs a variable quantization step size signal to keep the input image data and the output image data constant. . In other words, the output data in the buffer 14 is kept constant, but the input data varies with the amount of data depending on the amount of motion or complexity of the image. Therefore, the buffer supplies the quantization step size Q to the quantization unit 12 to convert the quantization level so as to adjust the amount of data generated by the first variable length encoding unit 13 so as to be within the buffer size. In the case where the buffer occupancy increases, that is, the cumulative data amount of the buffer increases, the quantization step size increases, and accordingly, the number of quantization levels decreases, so that the amount of data to be encoded is reduced and the input / output data amount of the buffer is kept constant.

In general, since there are many similar parts between the screen and the screen, the motion vector (MV) is calculated by estimating the motion in the case of a screen having a slight motion, and the data is compensated using the motion vector to determine the distance between the adjacent screens. Since the difference signal is very small, the transmission data can be further compressed. In order to perform such motion compensation, the inverse quantization unit 15 and the inverse DCT unit 16 inversely quantize the quantization coefficients output from the quantization unit 12 and inversely transform the image data of the spatial domain. The inversely transformed image data is reproduction error data corresponding to the error data output from the first adder A1. The screen is reconstructed by adding the error data output from the inverse DCT unit 16 to the block data fed back to the second adder A2 and storing it in the frame memory! 7. Then, the motion estimating unit 18 searches for block data having a pattern most similar to the block data input to the input terminal 10 from the frame data stored in the frame memory 17, and finds a motion vector MV representing the motion between two blocks. Calculate. This motion vector (MV) is transmitted to the receiving side via the second variable length coding unit 20 and to the motion compensating unit 19 at the same time for use in the decoding system. The motion compensator 19 reads block data corresponding to the motion vector from the frame data of the frame memory 17 according to the motion vector supplied from the motion estimation unit 18 and supplies it to the first adder A1. Then, as described above, the first adder calculates error data between the block data supplied from the input terminal 10 and the block data of the similar pattern supplied from the motion compensator 19, and the error data is encoded again. The encoded image data V _CD , the quantization step size Q, and the motion vector MV are thus combined in the data combiner 21 and transmitted to the receiving side as transmission data V _T.

The image data transmitted as described above is input to a decoder as shown in FIG. The data separator 22 of the decoder separates the transmission data V _T into the coded image data V _CD , the quantization step size Q, and the motion vector MV. Then, the encoded image data V _CD is decoded through a reverse process of variable length encoding by the first variable length decoding unit 24 through the buffer 23, and output from the first variable length decoding unit 24. The data is inversely quantized by the inverse quantization unit 25 by the conversion coefficient of the frequency band. The inverse DCT unit 26 converts the frequency domain transformation coefficient supplied from the inverse quantization unit 25 into image data of the spatial domain. In this case, the inverse transformed image data is reproduction error data corresponding to the error data calculated from the first adder A1 in the encoder of FIG. In addition, the motion vector (MV) calculated and transmitted by the motion estimation unit 18 of the encoder is similarly separated by the data separation unit 22 and decoded by the second variable length decoding unit 27, and then the motion compensation unit 29 The motion compensator 29 reads block data corresponding to the motion vector MV from the frame data stored in the frame memory 28 and supplies it to the adder A. FIG. Then, the inversely transformed error data and the block data supplied from the motion compensation unit 29 are combined in the adder A and transmitted to the display unit.

However, in the conventional system, when data compression of a video signal having a wide bandwidth, such as a video signal such as a high-definition TV (HD-TV), is subjected to severe band compression, quantization noise is added to the encoding and decoding process. As a result, the image data is damaged and the image quality deteriorates.

Accordingly, it is an object of the present invention to encode encoded image data in block units of a predetermined size, and to adaptively calculate predetermined coefficient data for correcting the degree of damage of the encoded and decoded image data. The present invention provides an encoding system for transmitting more efficient encoding data by outputting the same together.

Another object of the present invention is to further improve image quality by adaptively post-processing a decoded video signal using the predetermined coefficient data transmitted from the encoding system in decoding the image data transmitted from the encoding system as described above. In providing a decoding system for providing a.

An object of the present invention is to provide a video data encoding system for outputting encoded data, comprising means for encoding image data divided into blocks of a predetermined size and means for reconstructing quantized data for performing prediction encoding. By means of receiving the image data input to be encoded and the image data reconstructed by the reconstruction means, generating means and generating coefficient values for minimizing the difference between the two data.

Another object of the present invention is to provide a decoding system having means for receiving and decoding encoded image data divided into blocks having a predetermined size, the means for separating the predetermined coefficient data and the encoded image data, and supplied from the separation means. It is achieved by post-processing means for receiving the coefficient value and the decoded image data by the decoding means to correct the decoded image data by using the coefficient value.

Hereinafter, a preferred embodiment of an encoding and decoding system using adaptive processing according to the present invention will be described in detail with reference to the accompanying drawings.

3 is a block diagram illustrating an image data encoding apparatus according to an embodiment of the present invention, and illustrates an encoding apparatus for adaptively processing an image signal predicted and encoded in units of N × N blocks. In the apparatus of FIG. 3, the apparatus of FIG. 1 receives the data including the image data coming through the input terminal 10 and the quantization noise output from the second adder A2, and performs data processing by adaptive processing. Adaptive processing means for outputting a predetermined signal to the unit 21 is added.

The above-described adaptive processing means has a delay unit 31 for time-delaying the image data input through the input terminal 10, and its input terminal to the output terminal of the second adder A2. The adaptive processing unit 32 connected to the unit 21 and the output signal of the delay unit 31 are applied to the input terminal of one side, and the output signal of the adaptive processor 32 is applied to the input terminal of the other side. And a third adder A3 for generating a signal for controlling the adaptive processor 32. In FIG. 3, the same reference numerals are given to the same components as those in FIG. 1, and the data combining unit 21 to which predetermined coefficient data is applied is also assigned the same reference numerals. In addition, since the description of the operation of the same element is described above, it will be omitted.

The delay unit 31 time-delays the image data sampled in units of N × N blocks (generally 8 × 8 blocks) input through the input terminal 10 by a predetermined time interval. For example, the delay unit 31 encodes and decodes the K-th image data D _k sampled in the time domain while passing through the inverse DCT unit 16 from the DCT unit 11 and the second adder ( The above-described K-th image data D _k is time-delayed by a time interval output in the form of noise-containing image data X _k through a process of adding the block data fed back in A2). The K-th noise-containing image data X _k outputted after being delayed by the delay unit 31 is input to the adaptive processor 32. The adaptive processing unit 32 changes the quantization step size (Q) to the image data interval (hereinafter referred to as slice) of image data belonging to 16 horizontal scan lines, thereby minimizing the mean square error (Least). The adaptive coefficient vector (W _S ) for each slice is obtained by using a Mean Square Error method. Here, the subscript of S W _S shows that the adaptive coefficient vector W _S for a particular slice. When the adaptive processor 32 generates pseudo image data P _{k of} the current pixel using data correlation of five adjacent pixels, the third adder A3 generates the pseudo image data P _k . An error signal E _k corresponding to a difference between the image data and the original image data D _k delayed output from the delay unit 31 is generated and applied to the adaptive processor 32.

In order to minimize the above error signal E _k , the adaptive processor 32 uses the following equations (1) and (2).

Here, α is a coefficient for faster calculation of the adaptive coefficient vector W _S for minimizing the error signal E _k , which is a difference between the original image data D _k and the pseudo image data P _k . It is an experimental value having a range of 1. The pseudo adaptation coefficient vector W _k is an adaptation coefficient for five pixels having a mutually specific positional relationship adjacent to each other by using the K th pixel sampled in the time domain as the current pixel as a reference for operation W ⁰ _k , W ¹ _k , W ² _k , W ³ _k , and W ⁴ _k are represented by the components thereof. In addition, the noise-containing image data vector N _k represents noise-containing image data of five adjacent pixels having the above-mentioned K-th pixel as a reference pixel. The pseudo image data P _k of Equation (2) is pseudo image data for the K th pixel.

4 is a conceptual diagram illustrating a correlation between a current pixel and an adjacent pixel, and illustrates a relationship between the current pixel and four pixels adjacent thereto. In FIG. 4A, the spatial position of the current pixel is defined as a variable (m, n) as a two-dimensional coordinate system, and the position of the immediately preceding pixel on the same horizontal scan line as the current pixel is represented by (m-1, n). In addition, the position of the pixel before the horizontal scanning line to which the current pixel belongs is indicated by (m, n-1), and the position of the immediately preceding pixel on the same horizontal scanning line as the pixel at the position (m, n-1) is indicated. The position is represented by (m-1, n-1), and the position of the subsequent pixel is represented by (m + 1, n-1). 4B illustrates a process of multiplying the noise-containing image data X _{k of the} current pixel and the data of four pixels adjacent to the current pixel by multiplying corresponding pseudo adaptation coefficients.

The noise-containing image data (X _k ) input to the adaptive processor 32 is an adaptive coefficient (component) of the pseudo-adaptive coefficient vector (W _k ) to each of adjacent pixels as shown in FIG. 4 and equation (2). W ⁱ _k) the product made following the values P _k to a summed result (i.e., P _k = W ⁰ _k and X (m, n) + W ¹ _k and X (m-1, n) + W 2 k • X (m-1, n-1) + W ³ _k ㆍ X (m, n-1) + W ⁴ _k ㆍ Output as X (m = 1, n-1) In the adaptive processing unit 32 The output pseudo image data P _k and the original image data D _k delayed output from the delay unit 31 are input to the third adder A3, and the original image data from the third adder A3. D _k ) and the error signal E _k corresponding to the difference between the pseudo image data P _k is outputted and sent to the adaptive processor 32, the adaptive processor 32 using the equation (1) using the error signal. The component values of the adaptive coefficient vector W _k are updated in the direction of minimizing (E _k ), that is, the pseudo adaptation coefficient vector W _k for the K th pixel. One slice by repeating the method of obtaining the pseudo image data (P _{k + 1} ) for the K + 1 th pixel and comparing it with the K + 1 th original image data (D _{k + 1} ) using The optimal adaptive coefficient vector W _S is obtained for minimizing the error signal for all the pixels in the pixel, so that one adaptive coefficient vector W _S composed of five adaptive coefficients is obtained in one slice. After the estimation of the adaptive coefficient vector W _S for the data in the current slice is completed and the next slice is passed, the adaptive coefficient vector W _S for the slice is the same as the method of obtaining the above-described adaptive coefficient vector. Calculate the way.

The adaptive coefficient vector W _S for the current slice output from the adaptive processor 32 is applied to the data combiner 21. The data combiner 21 encodes the encoded image data V _CD outputted from the buffer 14 by components W ⁰ _S , W ¹ _S , W ² _S , W ³ _S , and W ⁴ _S of the adaptive coefficient vector. And the transmission vector V _{TR are} combined with the motion vector MV output from the quantization step size Q and the second variable length coding unit 20 to transmit the gen-transmission data V _TR to the receiver.

5 is a block diagram illustrating an image data decoding apparatus according to an embodiment of the present invention, and illustrates a decoding apparatus for post-processing adaptively processed image data. The apparatus of FIG. 5 adds the post processor 33 to the apparatus of FIG. 2 described above. One terminal of the post processor 33 is connected to the data separator 22 and the other terminal is decrypted as shown in FIG. Connect to the output of the device. In FIG. 5, the same reference numerals are given to the same components as those in FIG. 2, and the data separator 22 separating coefficient data is also given the same reference numerals. In addition, since the description of the operation of the same element is described above, it will be omitted.

The transmission data V _TR input to the data separator 22 is separated by the data separator 22 and component data W ⁰ _S ˜ of the adaptive coefficient vector W _S among the data generated by the data separator 22. W ⁴ _S ) is input to the post-processing section 33. Compensation according to the motion estimation of the image data output through the inverse quantization and inverse DCT transformation in the above-described encoding apparatus and the variable length encoding process are performed by inverse quantization and inverse DCT transformation in the decoding apparatus of FIG. It is the same as the process of generating motion compensated image data using. Therefore, the K-th image data in the time domain input to the other input terminal of the post-processing unit 33 is almost the same as the K-th image data including the quantization noise output from the second adder A2 of FIG. Equation (2) used to obtain the image data can be used. The post processor 33 calculates image data of the current pixel including the adaptive coefficients W ⁰ _S to W ⁴ _S and quantization noise and image data of four adjacent pixels by using the above-described equation (2). Process. That is, as shown in FIG. 4, the data X (m, n) containing the quantization noise for the current pixel at position (m, n) is multiplied by the adaptive coefficient W ⁰ _S and adjacent to the position (m-1, n). The data X (m-1, n) containing the quantization noise for the pixel is multiplied by the adaptation coefficient W ¹ _S. When the multiplication operation for the remaining three adjacent pixels is completed in the same manner, the five data values generated by the multiplication operation are summed. The data Y _k generated by the data summation is the data most similar to the K th data D _k among the original image data input to the above-described encoder, and is data that minimizes quantization noise included in the encoding and decoding process. The process of calculating reconstructed data of each pixel is repeated for one slice using the same adaptive coefficient vector as in Equation (2). The image data obtained by varying the adaptive coefficient vector for each slice is supplied to the image display unit and displayed as an image of improved quality.

In the encoding and decoding system of the present invention as described above, the image quality of the image can be improved by minimizing additional noise such as quantization noise included in the encoding and decoding process.

Claims (16)

A method of encoding image data for outputting encoded data, the method comprising: encoding image data divided into blocks having a predetermined size and restoring quantized data to perform predictive encoding. An image data input step of applying image data and image data before encoding; An adaptive processing step of generating predetermined coefficient data at a predetermined amount of image data intervals so that the reconstructed image data is most similar to the image data before encoding using the reconstructed image data and the image data before encoding; ; And outputting the encoded image data and predetermined coefficient data. 2. The method of claim 1, wherein the adaptive processing step comprises: delaying image data before being encoded for a predetermined time interval; Storing the restored image data; Generating predetermined coefficient data which minimizes a mean square error of the image data of the pixels using the correlation of the image data of the predetermined number of pixels spatially adjacent to each other. A video data encoding method. 2. The method of claim 1, wherein the adaptive processing step generates the predetermined coefficient data at intervals of image data corresponding to sixteen horizontal scan lines. 3. The method of claim 2, wherein the delaying step time delays the same data as the data before encoding during a time interval during which the data before encoding is output through a process of encoding and restoring. 4. The method of claim 2 or 3, wherein the predetermined coefficient data is for five pixels that are spatially adjacent to each other. 6. The method of claim 5, wherein the coefficient data is five coefficient data corresponding to five spatially adjacent pixels. CLAIMS 1. A video data decoding method comprising the steps of decoding video data divided and encoded into blocks having a predetermined size, comprising: inputting predetermined coefficient data and encoded video data; And reconstructing the predetermined coefficient data and the image data generated in the decoding step into data similar to the image data before encoding. 8. The method of claim 7, wherein the coefficient data is predetermined coefficient data such that the reconstructed image data for predictive encoding is most similar to the image data before encoding. A coding apparatus for video data, comprising: means for encoding image data divided into blocks of a predetermined size and means for restoring quantized data to perform predictive encoding, and outputting the encoded data, the image being input for encoding. Means for receiving the data and the image data reconstructed by the reconstruction means and generating predetermined coefficient data for minimizing the difference between the two data; And data combining means for combining and outputting the encoded image data and the predetermined coefficient data. 10. The apparatus of claim 9, wherein the coefficient generator comprises: a delay unit for delaying the video data before being encoded by a predetermined time interval; And an adaptive processor which receives the reconstructed image data and the data output from the delay unit and generates predetermined coefficient data by using a mean square error minimization method. The encoding device according to claim 9 or 10, wherein the coefficient generating means generates a predetermined coefficient by using data correlation of five adjacent pixels. 11. An encoding apparatus according to claim 9 or 10, wherein the coefficient generating means generates predetermined coefficients in image data units corresponding to sixteen horizontal scan lines. The method of claim 12, wherein the coefficient generating means generates error data corresponding to the difference between the pseudo image data generated by using the image data to be encoded and reconstructed, and the image data before encoding from the delay unit and to all pixels. And encoding five coefficient data by data units of 16 horizontal scan lines so as to minimize an error between pseudo image data and image data before encoding. CLAIMS 1. A decoding apparatus comprising means for receiving and decoding image data divided and encoded into blocks having a predetermined size, comprising: means for separating predetermined coefficient data and encoded image data; And post-processing means for receiving the coefficient data supplied from the separating means and the decoded image data and correcting the decoded image data using the coefficient data. 15. The decoding apparatus according to claim 14, wherein the post-processing means generates data most similar to the image data before encoding using the same coefficient data for the data corresponding to the 16 horizontal scan lines. 16. The decoding apparatus according to claim 15, wherein the coefficient data is the same coefficient data as predetermined coefficient data generated during encoding of image data corresponding to the reconstructed data for predictive encoding.