KR20040027047A - Encoding/decoding apparatus and method for image using predictive scanning - Google Patents

Abstract

PURPOSE: A device for coding/decoding an image using a predictive scanning method is provided to analyze scanning characteristics in a scanning mode generator by using predictive block information or predictive additional information, and to determine a scanning mode according to the analyzed characteristics, thereby reducing run size used for a variable length coding process. CONSTITUTION: A predictive block generator(52) predicts blocks to be currently coded, and generates a predictive block similar to a current block. An error block generator(54) calculates a difference between the generated predictive block and the current block, and generates an error block. A block converter(56) performs a predetermined operation in the error block, and converts the error block. A scanning unit(70) scans a converted block image. A scanning mode generator(60) determines a scanning method, and supplies the scanning method to the scanning unit(70). A variable length coder(80) variable length-codes a scanned image signal, and generates a bit stream.

Description

Image encoding / decoding apparatus and method for image using predictive scanning

The present invention relates to a method and apparatus for encoding and decoding using predictive scanning of an image signal. Specifically, the present invention relates to a method and apparatus for encoding and decoding image data for encoding and decoding an image by selecting a scanning method according to characteristics of a predicted block image used for encoding / decoding.

1 is a block diagram showing the configuration of a coding apparatus for encoding a conventional video signal. Most still or moving picture encoding apparatuses have the configuration as shown in FIG. 1, and the encoder divides an image into blocks having a predetermined size and performs an encoding process in units of blocks.

Referring to FIG. 1, a conventional encoder performs prediction on a block to be currently encoded using reference blocks to generate a prediction block similar to the current block, and between the generated prediction block and the current block. An error block generator 14 for calculating a difference and generating an error block, a block converter 16 for performing a predetermined operation on the error block and converting the same, a scanning unit 18 for scanning the converted block image, and scanning And a variable length encoder 20 for encoding the encoded video signal to generate a bitstream.

As an example of generating a prediction block by using the reference blocks in the prediction block generation unit 12 of the prior art, the current block is predicted from the previous image reproduced in the video and the current block from the neighboring blocks processed in the still image. There are cases where blocks are predicted.

First, referring to FIG. 2, a process of generating a prediction block in a video corresponds to a block 201 to be currently encoded in the current video 200 when the current video 200 is predicted using the previous video 202. The corresponding block 203 at the same position as the current block of the previous image 202 is set. Thereafter, the similar block 205 most similar to the current block 201 is found within the search interval 204 of the distance corresponding to the corresponding block 203. In this case, additional information for prediction is required. In the case of a video, motion information such as dx and dy is separately encoded as additional information and transmitted to the decoder.

Meanwhile, in the case of a still image, a prediction block corresponding to the currently encoded block is generated by using neighboring blocks that are already processed. A process of generating a prediction block of a still image will be described with reference to FIG. 3A.

FIG. 3A shows a prediction block corresponding to a current block composed of 4x4 pixels a to p, and reference pixels A to Q of already processed neighboring blocks used for generation of the prediction block.

There are various methods of setting each pixel value included in the prediction block using the reference pixels A to Q. A representative example is illustrated in FIGS. 3B to 3D.

3B illustrates a method of generating a predictive block using three modes from reference blocks. Referring to FIG. 3B, in mode 0, in the vertical direction prediction, all a, e, i, m pixels included in the prediction block are predicted by the value of pixel A of the reference block located on the same vertical line as these pixels. Similarly, the predicted values of pixels b, f, j, n are the values of pixel B, the predicted values of pixels c, g, k, o are the values of pixel C, and the predicted values of pixels d, h, i, p are pixels. Predict each of the values of D.

On the other hand, in the case of Mode 1, in the horizontal direction prediction, the method of setting the prediction value is the same as the above-described vertical direction prediction. Therefore, the prediction values set in the prediction block are on the same horizontal line among the values of I, J, K, and L pixels. Has

Finally, in mode 2, all pixels of the current block are predicted by the average of the peripheral reference pixels. 3C and 3D are diagrams illustrating examples of using more prediction directions shown in FIG. 3B, and show five modes and nine modes, respectively.

Prediction in still images also requires additional information, such as prediction in video, where the information on the prediction direction is additional information of the still image prediction, and information about the prediction direction is separately encoded and transmitted to the decoder.

Referring back to FIG. 1, the block transform unit performs a transform such as a DCT transform on an error image block to generate an image block composed of transform operation coefficients, and the converted block is finally scanned by the scanning unit 18 for variable length encoding. Is scanned.

In a standard related to conventional image compression, zigzag scanning as shown in FIGS. 4A and 4B is used. 4A illustrates a zigzag scanning of a 4 × 4 block, and FIG. 4B illustrates a zigzag scanning of a 8 × 8 block. Through this scanning, a non-zero coefficient and a run of zeros immediately before the coefficient are calculated, and the image block is encoded by encoding the level and the run. FIG. 4C illustrates run and level calculated in the order shown in FIG. 4A in a 4 × 4 image block.

In encoding the image according to the above-described method, the encoding efficiency is determined according to the length of the run included in the signal input to the variable length encoder 20. In other words, the smaller the size of the run, the smaller the number of bits allocated to the encoding of the run, and the smaller the number of bits required for the entire encoding. Therefore, the encoding efficiency is increased. Since the size of the run is determined according to the method of scanning the image block output from the block converter 16 in the scanning unit 18, a scanning method for minimizing the size of the run is required.

As a method of encoding an image by minimizing the size of the run, there is a method of increasing compression efficiency by selectively using the existing zigzag scanning method and another scanning method according to the situation.

However, this method uses a predefined scanning method in addition to the zigzag scanning method, so that both the zigzag scanning method and the predefined scanning method are inefficient. In addition, since the information on the selected scanning method must be encoded and provided to the decoder, additional bits are generated. Currently, this method has been selected as the scanning method of the MPEG-4 intra coding method, but there is a problem that a lot of bits occur rather than using only the zigzag scanning method.

As another method of encoding an image by minimizing the size of the run, there is a method of presetting a plurality of scanning methods in advance than two scanning methods to select the most suitable scanning method and applying the same to the image blocks.

However, since this method defines a plurality of scanning methods in advance, it may be an inefficient scanning method depending on the type of data. That is, when a plurality of scanning methods are defined, the information on the scanning method is so large that the overall amount of bits is generated, and thus the predefined finite scanning method may not be effective for all data. This is an essential problem in the basic structure that must define a finite scanning method and provide information on the selected scanning method.

SUMMARY OF THE INVENTION The first technical problem to be solved by the present invention is to provide an encoding method and an encoding apparatus which improves encoding efficiency and does not require encoding of additional information about a scanning method.

Another object of the present invention is to provide a decoding apparatus and method for decoding image data encoded by the encoding apparatus of the present invention.

1 is a block diagram showing the structure of a conventional predictive encoder.

2 is a diagram illustrating a motion prediction method in a video.

3A to 3D are diagrams for describing a method of predicting a current block from adjacent pixels in a still image.

4 is a diagram illustrating a conventional zigzag scanning structure and an example of zigzag scanning.

5A is a block diagram showing the configuration of a predictive encoder according to a first preferred embodiment of the present invention, and FIG. 5B is a flowchart illustrating a predictive coding method according to the first preferred embodiment of the present invention.

FIG. 6A is a block diagram illustrating a configuration of the first scanning mode generator illustrated in FIG. 5A, and FIG. 6B is a flowchart illustrating a method of generating the first scanning mode.

FIG. 7 is a block diagram showing the configuration of the scanning unit shown in FIG. 5A.

8 is a diagram illustrating a scanning sequence according to the scanning mode of the present invention.

FIG. 9A is a diagram illustrating a prediction image composed of prediction blocks, and FIG. 9B is a diagram illustrating an error image composed of error blocks.

10 is a diagram illustrating an example of distribution of transform coefficients in accordance with a prediction direction.

FIG. 11A is a block diagram showing the configuration of a predictive encoder according to the second preferred embodiment of the present invention, and FIG. 11B is a block diagram showing the configuration of the second scanning mode generator shown in FIG. 11A.

12A is a block diagram showing a configuration of a decoding apparatus according to a first embodiment of the present invention, and FIG. 12B is a flowchart for explaining a decoding method according to the first embodiment.

13 is a block diagram showing the configuration of a decoding apparatus according to a second preferred embodiment of the present invention.

14 is a diagram comparing the performance of the conventional zigzag scanning method and the predictive scanning method of the present invention.

According to a first aspect of the present invention, there is provided a prediction encoding apparatus comprising: a prediction block generation unit generating a prediction block for a current block to be encoded from base-coded reference blocks; A block transform unit configured to generate a transform block composed of transformed values by performing a predetermined transform operation on an error block between a current block and a prediction block; A scanning mode generator that determines a scanning mode by analyzing edge directions of an image included in the prediction block; A scanning unit scanning a transform block according to a scanning mode; And an entropy encoder for encoding the image data input from the scanning unit.

According to a first aspect of the present invention, there is provided a prediction encoding method comprising: (a) generating a prediction block for a current block to be encoded from base coded reference blocks; (b) generating a transform block composed of the transformed values by performing a predetermined transform operation on an error block between the current block and the prediction block; (c) determining a scanning mode by analyzing edge directions of an image included in the prediction block; (d) scanning the transform block according to the scanning mode; And (e) entropy encoding the scanned image data.

The prediction encoding apparatus according to the second embodiment of the present invention for achieving the first technical problem described above generates a prediction block for a current block to be encoded from base coded reference blocks according to a predetermined prediction direction. A prediction block generator; A block transform unit configured to generate a transform block composed of transformed values by performing a predetermined transform operation on an error block between a current block and a prediction block; A scanning mode generator that determines a scanning mode according to a predetermined prediction direction; A scanning unit scanning a transform block according to a scanning mode; And an entropy encoder for encoding the image data input from the scanning unit.

The prediction encoding method according to the second embodiment of the present invention for achieving the first technical problem of the present invention described above includes (a) a prediction block for a current block to be encoded from base coded reference blocks according to a predetermined prediction direction. Generating a; (b) generating a transform block composed of the transformed values by performing a predetermined transform operation on an error block between the current block and the prediction block; (c) determining a scanning mode according to a predetermined prediction direction; (d) scanning the transform block according to the scanning mode; And (e) entropy encoding the scanned image data.

According to a second aspect of the present invention, there is provided a predictive decoding apparatus, including: a variable length decoder configured to decode an input bitstream to generate a data sequence; A prediction compensator for decoding the prediction side information from the input bitstream to generate a prediction block from the reconstructed reference blocks; A scanning mode generator for detecting an edge of the prediction block and determining an edge direction to generate a scanning mode; An inverse scanning unit scanning the decoded data stream according to the scanning mode to generate a transform block; A block inverse transform unit generating an error block by performing a predetermined operation on the transform block; And a block reconstruction unit that generates a reconstructed image block by adding the prediction blocks input from the prediction compensator to the error blocks.

According to a second aspect of the present invention, there is provided a predictive decoding method including (a) generating a data string by decoding an input bitstream; (b) decoding the prediction side information from the input bitstream to generate a prediction block from the reconstructed reference blocks; (c) generating a scanning mode by detecting an edge of the prediction block and determining an edge direction; (d) scanning the decrypted data stream according to the scanning mode to generate a transform block; (e) generating an error block by performing a predetermined operation on the transform block; And (f) generating the reconstructed image block by adding the prediction block to the error block.

Hereinafter, a scanning method and an encoding apparatus of the present invention will be described with reference to the accompanying drawings.

FIG. 5A is a block diagram showing a configuration of an encoding apparatus according to a first preferred embodiment of the present invention, and FIG. 5B is a flowchart for explaining an encoding method according to the first embodiment of the present invention.

Referring to FIG. 5A, the encoding apparatus according to the first embodiment of the present invention performs a prediction on a block to be currently encoded using reference blocks to generate a prediction block similar to the current block. The error block generator 54 generates an error block by calculating a difference between the generated prediction block and the current block, the block converter 56 converting the error block by performing a predetermined operation, and scans the converted block image. The scanning unit 70 to determine the scanning method according to the edge detected by the prediction block, the scanning mode generator 60 to provide the scanning unit, and the variable length encoding to encode the scanned image signal to generate a bitstream. The encoder 80 is included.

Referring to FIG. 5B, when the encoding method of the present invention is described, when a block to be currently encoded is input to the encoder of the present invention, the prediction block generator 52 of the present invention refers to a reference adjacent to the current block in the case of a still image. In the case of a video, blocks are used to generate a prediction block using reference blocks of a previous frame (S500).

The error block generator 54 receives the generated prediction block and the current block, calculates an error between the current block and the prediction block, generates an error block, and outputs the generated error block to the block converter 56 ( S510).

The block converter 56 performs a predetermined operation on the input error block, and generates a converted image block composed of the result of the operation. In the preferred embodiment of the present invention, the block converter 56 is applied to the error block. The DCT operation is performed and the transform block including the DCT coefficient value is output to the scanning unit 70 (S520).

Meanwhile, the first scanning mode generator 60 receives the prediction block output from the prediction block generator 52, detects the edge and the direction of the prediction block, and determines a scanning method to be performed on the transform block. Output to 70 (S530).

The scanning unit 70 receiving the transform block and the scanning mode configured with the DCT coefficients outputs the data string generated by scanning and scanning the transform block in accordance with the input scanning mode (S540).

The variable length encoder 80 encodes the input data string using the level and the run included in the data string input from the scanning unit 70 (S550).

When the scanning mode generator 60 of the present invention performs scanning in a direction perpendicular to the edges included in the prediction block, the run that determines the size of the variable length coded data is minimized. Determine the scanning mode. In addition, the scanning mode generator of the present invention determines the scanning mode by using similarity in the distribution of data of the error block and the transform block to be variable-length coded.

Most of the prediction errors that occur during the prediction process are centered around the edges of the image, and most of the edges are directional. That is, the prediction error is distributed with directionality, and when the error image having the directionality is transformed, the transformed coefficient value also shows a different distribution depending on the direction of the edge.

9A illustrates an image predicted through prediction in a still image, and FIG. 9B illustrates a prediction error image. Referring to FIG. 9B, it can be seen that most of the errors appear strongly around the edges, and therefore, the errors are distributed with directionality. This phenomenon also occurs in motion estimation and prediction through compensation.

Therefore, when scanning is performed in a direction orthogonal to the edge of the transform block having the directionality similar to that of the edge, the size of the run is reduced and the coding efficiency can be improved.

6A and 6B, a first scanning mode generator 60 according to a first embodiment of the present invention will be described.

FIG. 6A is a block diagram illustrating a configuration of the first scanning mode generator 60 of the present invention. The first scanning mode generator 60 includes an edge detector 62 and an edge detector for detecting an edge from an input prediction block. A direction determining unit 64 determines a direction of the detected edge, and a first mode determining unit 66 that determines a scanning direction according to the direction of the edge determined by the direction detecting unit to generate a scanning mode.

Referring to FIG. 6B, which is a flowchart illustrating a scanning mode generation method, first, the edge detector 62 receives a prediction block generated from the prediction block generator (S532) and includes the prediction block in a prediction block according to various edge detection methods. The detected edge is detected (S534).

Edge detection can be determined by using sobel operator, canny edge detection, laplacian operator, etc. and applying it to the threshold. For example, when the edge detector 62 detects an edge using a convolution operation, the edge detector 62 performs a convolution operation while moving a 3 × 3 mask on the prediction block. If the result of the convolution operation is equal to or greater than a predetermined threshold, the corresponding pixel is determined to belong to the edge.

On the other hand, the edge detected by the edge detection unit 62 is input to the direction determiner 64, the direction determiner 64 analyzes the distribution of the pixels constituting the detected edge by the x-axis and y-axis components The direction of the edge is determined (S536).

For example, the sum of the x-coordinate value and the y-coordinate value is compared with respect to the position (x, y) of the edge pixels. When the sum of the x-coordinate values is large, the direction is horizontal. You can judge in the direction. In addition, when the sum of the x and y coordinate values is large and there is little difference, the diagonal direction is determined.

The first mode determiner 66 determines a scanning mode to be performed on the converted block according to the edge direction input from the direction determiner 64 (S538). When the edge is horizontal, large coefficient values are distributed in the vertical direction because of the characteristics of the transform, and even if quantized, non-zero coefficient values are distributed in the vertical direction. Therefore, in this case, scanning can be performed by scanning to have a vertical tendency rather than zigzag scanning.

When the edge is vertical, it is effective to scan such that the direction is horizontal as opposed to the horizontal case. In addition, when the edges are distributed diagonally, it is effective to allow the coefficients on the diagonal to be scanned before the zigzag scanning.

7 is a block diagram showing the configuration of a scanning unit of the present invention.

Referring to FIG. 7, a scanning unit 70 according to a preferred embodiment of the present invention includes a scanning processor 72, a selector 74, and a scanning order storage unit 76.

When the scanning mode is input to the selector 74 of the scanning unit, the selector 74 selects one of the scanning orders recorded in the scanning order storage unit 76 according to the input scanning mode, and selects the selected scanning order into the scanning processing unit 72. ) The scanning processor 72 scans the transform block according to the input scanning order, generates a level and a run, and outputs the level and the run to the variable length encoder.

FIG. 8 is a diagram illustrating a scanning sequence corresponding to each mode when the scanning modes are 0 to 8. FIG.

Hereinafter, an encoding apparatus according to a second preferred embodiment will be described with reference to FIG. 11A.

The encoding apparatus according to the second embodiment includes a prediction block generator 110, an error block generator 112, a block converter 114, a second scanning mode generator 120, a scanning unit 130, and a variable length. The encoder 140 is included. Since the configuration of the encoding apparatus according to the second embodiment is the same as the configuration of the encoding apparatus according to the first embodiment except for the second scanning mode generator 120, a diagram illustrating the configuration of the second scanning mode generator will be described below. Only the second scanning mode generator will be described with reference to 11b.

Referring to FIG. 11B, the second scanning mode generator of the present invention includes the prediction direction determiner 122 and the second mode determiner 124 to generate additional information (prediction direction or prediction size) generated during the prediction process. Directly determine the prediction direction of the current block, and determine the scanning mode.

When additional information generated when the prediction block is generated is input to the prediction direction determiner 122 included in the second scanning mode generator 120, the prediction direction determiner 122 determines the prediction direction according to the input additional information. Output to the second mode determiner 124. The prediction direction determiner 122 determines the direction shown in FIGS. 3B to 3D as the prediction direction in the case of a still image, generates a motion vector as shown in FIG. 2 in the case of a video, and The components can be analyzed to determine the prediction direction.

The second mode determiner 124 determines the scanning mode according to the prediction direction input from the prediction direction determiner 122, and outputs the scanning mode to the scanning unit 130. The determination method is the first mode determiner 66 described above. Is similar to).

Since the direction of the edge detected by the edge detector of the first embodiment is generally similar to the prediction direction, the direction of the scanning mode determined by the second mode determiner 124 represents a direction having a property perpendicular to the prediction direction. In the case of the first embodiment, the same scanning mode is generated.

For example, in the case of FIG. 3A, when the prediction block is generated in the vertical direction as in mode 0, vertical edges are mainly generated. In addition, when the prediction block is generated in the horizontal direction as in Mode 1, edges in the horizontal direction are mainly generated. Therefore, since the prediction scanning method of the present invention increases the coding efficiency by performing scanning in a direction perpendicular to the edge direction of the prediction block, the scanning mode determined by the second mode determiner performs scanning in a direction perpendicular to the prediction direction. And the result is the same as the scanning mode generated by the first scanning mode generator.

Meanwhile, in the case of the second scanning mode generator 120 described above, the scanning mode may be determined by only determining the prediction direction using additional information generated in the prediction process. The process is simpler than the first scanning mode generator which requires the process of determining the direction.

10 shows that the current block is classified by the prediction direction information generated in the prediction process by using a sequence called foreman. The absolute value of the coefficient values is accumulated for all blocks of 30 sheets. 10 shows that the transformed coefficient values are distributed differently according to the direction of each edge. Accordingly, the present invention can perform more effective scanning by using the distribution of the change coefficient values for the edge direction well.

FIG. 14 shows the results of experiments of the scanning method used in the predictive coding apparatus of the present invention and the conventional zigzag scanning method for various sequences. 14 shows the total number of runs for each chapter for each sequence. The more effective scanning is, the smaller the value of run is. Therefore, in FIG. 14, it can be seen that all show effective scanning performance as compared to the conventional zigzag scanning.

Hereinafter, a decoding apparatus and a method for reconstructing an image block by decoding a coded bitstream in the prediction encoding apparatus according to the first embodiment of the present invention will be described with reference to FIGS. 12A and 12B.

Referring to FIG. 12A, which is a block diagram illustrating a configuration of a decoding apparatus, a decoding apparatus includes a variable length decoder 150, a predictive compensator 152, a third scanning mode generator 154, an inverse scanning unit 156, and a block. An inverse transform unit 158 and a block reconstruction unit 160.

Referring to FIG. 12B, which is a flowchart illustrating a decoding method, the variable length decoder 150 generates a data string by decoding an input bitstream, and outputs the generated data string to the inverse scanning unit 156 (S1200). ).

The prediction compensator 152 decodes the prediction side information from the input bitstream to generate a prediction block from the already reconstructed reference blocks, and outputs the generated prediction block to the third scanning mode generator 154 (S1220).

The third scanning mode generator 154 detects the edge of the prediction block, determines the edge direction, generates the scanning mode in a direction orthogonal to the edge direction, and performs the reverse scanning unit ( 156) (S1240).

The inverse scanning unit 156 inversely scans the decoded data string input from the variable length decoder 150 according to the scanning mode to generate a transform block, and outputs the generated transform block to the block inverse transform unit 158 ( S1260).

The block inverse transform unit 158 generates an error block by performing an inverse transform of the transform used for predictive encoding, and outputs the generated error block to the block reconstruction unit 160 (S1280).

The block reconstruction unit 160 outputs the reconstructed image block by adding the prediction blocks input from the prediction compensator 152 to the error blocks input from the block inverse transform unit 158 (S1300).

FIG. 13 is a block diagram illustrating a configuration of a decoding apparatus for reconstructing an image block by decoding a coded bitstream in the prediction encoding apparatus according to the second embodiment of the present invention.

The configuration of the decoding apparatus illustrated in FIG. 13 is the same as that of the decoding apparatus illustrated in FIG. 12A except for the fourth scanning mode generator 174. In addition, the operation of the fourth scanning mode generator 174 is the same as the second scanning mode generator 120 of the prediction encoding apparatus according to the second embodiment of the present invention described above.

The fourth scanning mode generator 174 determines the prediction direction according to the prediction side information decoded from the input bitstream, determines the scanning mode according to the determined prediction direction, and outputs the scanning mode to the inverse scanning unit 176. Is similar to the case of the first scanning mode generator of the prediction encoding apparatus described above.

The invention can also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like, which are also implemented in the form of a carrier wave (for example, transmission over the Internet). It also includes. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

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 appended claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

The prediction encoding apparatus of the present invention analyzes a characteristic for scanning in a scanning mode generator using a predictive block or predictive additional information, and determines a scanning mode according to the analyzed characteristic to determine a length of a run used for variable length encoding. It is possible to improve the coding efficiency by reducing the size.

In addition, since the prediction encoding / decoding apparatus of the present invention determines the scanning method according to the prediction block or the prediction additional information, it is not necessary to encode information indicating a separate scanning method, thereby improving the coding efficiency.

Claims (22)

A predictive encoding device for encoding an image in block units, A prediction block generation unit generating a prediction block for a current block to be encoded from the base coded reference blocks; A block transform unit configured to generate a transform block composed of transformed values by performing a predetermined transform operation on an error block between the current block and the prediction block; A scanning mode generator that determines a scanning mode by analyzing edge directions of an image included in the prediction block; A scanning unit scanning the transform block according to the scanning mode; And And an entropy encoder for encoding the image data input from the scanning unit. The method of claim 1, wherein the scanning mode generator An edge detector detecting an edge included in the image of the prediction block; A direction determining unit which determines a direction component of the detected edge; And And a mode determiner configured to determine a scanning direction according to the determined edge direction to generate a scanning mode. The method of claim 2, wherein the mode determination unit And the scanning direction is determined such that the direction of the edge and the scanning direction are orthogonal to each other. The method of claim 1, The entropy encoder is a variable length encoder. A predictive encoding device for encoding an image in block units, A prediction block generator for generating a prediction block for the current block to be encoded from the base coded reference blocks according to a predetermined prediction direction; A block transform unit configured to generate a transform block composed of transformed values by performing a predetermined transform operation on an error block between the current block and the prediction block; A scanning mode generator that determines a scanning mode according to the predetermined prediction direction; A scanning unit scanning the transform block according to the scanning mode; And And an entropy encoder for encoding the image data input from the scanning unit. The method of claim 5, wherein the scanning mode generator A prediction direction determiner which determines a prediction direction used to generate a prediction block corresponding to the current block; And And a mode determiner configured to determine a scanning mode according to the prediction direction. The method of claim 6, wherein the prediction direction determiner If the prediction block is included in a previous frame of the frame including the current block, the encoding device is characterized by determining the prediction direction by analyzing the components of the motion vector moved from the block corresponding to the current block . The method of claim 6, wherein the mode determination unit And a scanning mode is determined such that the prediction direction and the scanning direction are orthogonal to each other. The method of claim 5, wherein And the entropy encoder is a variable length encoder. A prediction encoding method of encoding an image in block units, (a) generating a prediction block for the current block to be encoded from the base coded reference blocks; (b) generating a transform block composed of transformed values by performing a predetermined transform operation on an error block between the current block and the prediction block; (c) determining a scanning mode by analyzing edge directions of an image included in the prediction block; (d) scanning the transform block in accordance with the scanning mode; And (e) entropy encoding the scanned image data. The method of claim 10, wherein step (c) (c1) detecting an edge included in the image of the prediction block; (c2) determining a directional component of the detected edge; And (c3) generating a scanning mode by determining a scanning direction according to the determined edge direction. The method of claim 11, wherein step (c3) And the scanning direction is determined such that the direction of the edge and the scanning direction are orthogonal to each other. The method of claim 10, wherein step (e) A predictive encoding method characterized by variable length encoding the scanned image data. A prediction encoding method of encoding an image in block units, (a) generating a prediction block for a current block to be encoded from base coded reference blocks according to a predetermined prediction direction; (b) generating a transform block composed of transformed values by performing a predetermined transform operation on an error block between the current block and the prediction block; (c) determining a scanning mode according to the predetermined prediction direction; (d) scanning the transform block in accordance with the scanning mode; And (e) entropy encoding the scanned image data. 15. The method of claim 14, wherein step (c) (c1) determining a prediction direction used to generate a prediction block corresponding to the current block; And (c2) determining a scanning mode according to the prediction direction. The method of claim 15, wherein step (c1) When the prediction block is included in a previous frame of the frame including the current block, the prediction block is determined by analyzing the components of the motion vector moved from the block corresponding to the current block to determine the prediction direction Way. The method of claim 15, wherein step (c2) And a scanning mode is determined such that the prediction direction and the scanning direction are orthogonal to each other. The method of claim 14, wherein step (e) A predictive encoding method characterized by variable length encoding the scanned image data. 19. A recording medium in which the predictive encoding method according to any one of claims 10 to 18 is readable by a computer and recorded with executable program code. A variable length decoder for decoding an input bitstream to generate a data string; A prediction compensator for decoding prediction side information from the input bitstream to generate a prediction block from reconstructed reference blocks; A scanning mode generator for detecting an edge of the prediction block and determining an edge direction to generate a scanning mode; An inverse scanning unit scanning the decoded data stream according to the scanning mode to generate a transform block; A block inverse transform unit generating an error block by performing a predetermined operation on the transform block; And And a block reconstruction unit configured to generate a reconstructed image block by adding the prediction block input from the prediction compensator to the error block. The method of claim 20, wherein the scanning mode generator And determining a prediction direction according to the prediction additional information decoded from the input bitstream, and generating a scanning mode according to the determined prediction direction and outputting the scanning mode to the inverse scanning unit. (a) decoding the input bitstream to generate a data sequence; (b) decoding a prediction side information from the input bitstream to generate a prediction block from the reconstructed reference blocks; (c) generating a scanning mode by detecting an edge of the prediction block and determining an edge direction; (d) scanning the decoded data stream according to the scanning mode to generate a transform block; (e) generating an error block by performing a predetermined operation on the transform block; And and generating a reconstructed image block by adding the prediction block to the error block.