KR20040028318A - Image encoding and decoding method and apparatus using spatial predictive coding - Google Patents

Abstract

PURPOSE: A device and a method for coding and decoding an image by using spatial predictive coding are provided to use only pixels of the upper block and the left block of the current block in 4.times.4 spatial predictive coding, thereby reducing an amount of calculation and enabling simultaneous processing with an upper right or lower left block. CONSTITUTION: A temporal/spatial predictive coding unit(100) performs spatial predictive coding for an intra-image and temporal predictive coding for an inter-image. When the temporal/spatial predictive coding unit(100) performs the spatial predictive coding, a luminance component and a chronimance component are respectively predictively coded and the prediction is performed using values of pixels in adjacent blocks. A conversion/quantization unit(101) performs conversion for predictively coded information through DCT(Discrete Cosine Transform) and the like and quantizes the converted predictively coded information, to perform loss compression for the quantized information. And an entropy coding unit(102) performs entropy coding for the quantized information to generate a bit string.

Description

Image encoding and decoding method and apparatus using spatial predictive coding

The present invention relates to encoding and decoding of an image, and in particular, to encoding using a spatial prediction method.

In the existing international standard technology for video compression, one image is encoded by dividing it into luminance and color components. First, one image is encoded without referring to another image, and then motion images are predicted later in time with reference to this image. And predictive encoding using a compensation method. An image obtained by encoding an image without referring to another image is an intra image, and an image using a moving prediction and compensation method by referring to another image is called an inter image. Both intra and inter images are lossy compressed via DCT, quantization, and entropy coding. At this time, since the intra image is not predicted in time, a spatial prediction method is used to increase the compression efficiency.

In ISO / IEC MPEG-4 and ITU-T H.263 + video compression technology, when performing spatial prediction on intra video, set blocks of 8x8 pixel units, DCT / quantize each block, and then surround DC and AC values. The compression efficiency is increased by predictive encoding by referring to the DC value and the AC value of the block.

Recently, ISO / IEC MPEG and ITU-T VCEG jointly formed a joint video team (JVT) to develop a new video coding standard. The JVT final committee recommends a technique for compressing using spatial prediction method when encoding intra video. This is included. In this technique, the spatial prediction method uses different prediction methods for the luminance component and the color component. The luminance component uses blocks of 4x4 or 16x16 size, and for 4x4, there are 9 prediction methods depending on the prediction direction. In this case, there are four prediction methods.

In the following, we will look at the 4x4 and 16x16 block prediction methods used in the JVT Final Committee Recommendations and how to encode each of them.

First, in the case of using a 4x4 block, as shown in FIG. 1A, eight prediction methods are used according to eight directions, and DC prediction methods are added to the nine prediction methods. For prediction, the values of the pixels of the neighboring blocks are used. These pixels are predicted using all 17 pixels as shown in FIG. 1B. Here, the numbers in FIG. 1A represent each method by a number, and the DC prediction method is method 2.

The numbers, names, and prediction methods for each method are as follows.

Method 0: vertical prediction

Only used when ABCD is referenceable.

a, e, i, m is A,

b, f, j, n is B,

c, g, k, o is C,

d, h, l, p are predicted using D.

Method 1: horizontal prediction

Used only when IJKL is referenceable.

a, b, c, d is I,

e, f, g, h are J,

i, j, k, l is K,

m, n, o, p are predicted using L.

Method 2: DC prediction

All pixel values from a to p

When A B C D and I J K L are all referenceable (A + B + C + D + I + J + K + L + 4) >> 3,

If only A B C D can be referenced, (A + B + C + D + 2) >> 2,

If only I J K L can be referenced, it is predicted using (I + J + K + L + 2) >> 2.

Method 3: diagonal down / left prediction

a is (A + 2B + C + I + 2J + K + 4) >> 3

b, e is (B + 2C + D + J + 2K + L + 4) >> 3

c, f, i are (C + 2D + E + K + 2L + M + 4) >> 3

d, g, j, m are (D + 2E + F + L + 2M + N + 4) >> 3

h, k, n are (E + 2F + G + M + 2N + O + 4) >> 3

l, o is (F + 2G + H + N + 2O + P + 4) >> 3

p is predicted using (G + H + O + P + 2) >> 2.

Method 4: diagonal down / right prediction

m is (J + 2K + L + 2) >> 2

i, n is (I + 2J + K + 2) >> 2

e, j, o is (Q + 2I + J + 2) >> 2

a, f, k, p are (A + 2Q + I + 2) >> 2

b, g, l are (Q + 2A + B + 2) >> 2

c, h is (A + 2B + C + 2) >> 2

d is predicted using (B + 2C + D + 2) >> 2.

Method 5: vertical-right prediction

a, j is (Q + A + 1) >> 1

b, k is (A + B + 1) >> 1

c, l is (B + C + 1) >> 1

d is (C + D + 1) >> 1

e, n is (I + 2Q + A + 2) >> 2

f, o is (Q + 2A + B + 2) >> 2

g, p is (A + 2B + C + 2) >> 2

h is (B + 2C + D + 2) >> 2

i is (Q + 2I + J + 2) >> 2

m is predicted using (I + 2J + K + 2) >> 2.

Method 6: horizontal-down prediction

a, g is (Q + I + 1) >> 1

b, h is (I + 2Q + A + 2) >> 2

c is (Q + 2A + B + 2) >> 2

d is (A + 2B + C + 2) >> 2

e, k is (I + J + 1) >> 1

f, l is (Q + 2I + J + 2) >> 2

i, o is (J + K + 1) >> 1

j, p is (I + 2J + K + 2) >> 2

m is (K + L + 1) >> 1

n is predicted using (J + 2K + L + 2) >> 2.

Method 7: vertical-left prediction

a is (2A + 2B + J + 2K + L + 4) >> 3

b, i is (B + C + 1) >> 1

c, j is (C + D + 1) >> 1

d, k is (D + E + 1) >> 1

l is (E + F + 1) >> 1

e is (A + 2B + C + K + 2L + M + 4) >> 3

f, m is (B + 2C + D + 2) >> 2

g, n is (C + 2D + E + 2) >> 2

h, o is (D + 2E + F + 2) >> 2

p is predicted using (E + 2F + G + 2) >> 2.

Method 8: horizontal-up prediction

a is (B + 2C + D + 2I + 2J + 4) >> 3

b is (C + 2D + E + I + 2J + K + 4) >> 3

c, e is (J + K + 1) >> 1

d, f is (J + 2K + L + 2) >> 2

g, i is (K + L + 1) >> 1

h, j is (K + 2L + M + 2) >> 2

l, n is (L + 2M + N + 2) >> 2

k, m is (L + M + 1) >> 1

o is (M + N + 1) >> 1

p is predicted using (M + 2N + O + 2) >> 2.

">>" here means a left shift operation. Methods 3 through 8 are available when A, B, C, D, I, J, K, L, Q are referenceable. If E, F, G, and H are not referenceable, D is used instead. If M, N, O, and P are not referenceable, L is used instead.

In addition, when looking at the 16x16 block size prediction method, there are four DC, vertical, horizontal, and plane prediction methods, and the first three are the same as those used in 4x4, but the block size is larger. It is a method of predicting the change of the value of the current block by calculating the slope of the horizontal and vertical directions with reference to the pixel values of the block, and constructing a plane equation with this value.

The 4x4 or 16x16 prediction method encodes and transmits to the bit string which method is used in units of 4x4 or 16x16 blocks, respectively, and decodes this value during decoding to perform prediction compensation in the same manner. In order to encode this value efficiently, the JVT final committee recommends the value of the prediction method of the above block and the left block in case of 4x4. Refer to the smaller of the two values and refer to the 1 bit if the current block uses the value. 1 "is encoded to indicate that the method is used, otherwise" 0 "is encoded to the same 1 bit to indicate that the method is not used, and 3 bits of fixed-length code are encoded to determine which method is used. Inform.

In this case, if the above block or the left block is not encoded or cannot be referred to by the intra method, the value to be referred to is the DC method. If the above block and the left block are encoded by the 16x16 intra method, the corresponding block is considered to use the DC method.

In the case of 16x16, (Imode, nc, AC) are grouped together and encoded into a variable length code, where Imode represents a 16x16 prediction method. nc means 0 means that the color component is 0, 1 means that there is only DC in the color component, and 2 means that both DC and AC are present. AC indicates whether or not there is an AC component in the 16x16 luminance component block, and 0 indicates that there is no and 1 indicates that there is. This value is encoded as a macroblock type of an intra image. If the macroblock type is 0, a 4x4 block prediction method is used, and the values from 1 to 24 are determined according to the value of (Imode, nc, AC). The smaller the value of mb_type, the shorter the variable length code is used, and the larger the mb_type value is, the longer the variable length code is used.

The present invention aims to reduce the amount of computation by limiting the pixels of a neighboring block to be directly above or to the left of the current block for prediction when spatial prediction coding is performed on an intra image, but predicted in units of 4x4 blocks. When processing a block is to be processed simultaneously with the block on the upper right or lower left of the current block.

Another technical problem to be achieved by the present invention is to increase coding efficiency by efficiently encoding based on statistical characteristics when predicting by 4x4 or 16x16 block units and encoding a prediction method by 4x4 or 16x16 block units.

Another object of the present invention is to provide a recording medium recorded with a program code executable by a computer in an encoding and decoding method using the above method.

FIG. 1A is a diagram illustrating eight prediction directions and values of a prediction method according to each direction used in a 4x4 block-based spatial prediction coding method in the JVT final committee recommendation.

FIG. 1B is a diagram showing pixels of neighboring blocks used for predicting the value of a current block when performing 4x4 block-wise spatial prediction coding in the JVT final committee recommendation.

FIG. 2 is a table used for variable length coding a prediction method when performing 16x16 block-wise prediction coding in a JVT final committee recommendation.

3A is a block diagram schematically illustrating an embodiment of an image encoding apparatus according to the present invention.

3B is a block diagram schematically illustrating an embodiment of an image decoding apparatus according to the present invention.

4A is a flowchart illustrating an embodiment of an image encoding method performed by the apparatus of FIG. 3A.

4B is a flowchart illustrating an embodiment of an image decoding method performed in the apparatus of FIG. 3B.

FIG. 5 is a diagram illustrating pixels of neighboring blocks used to predict a value of a current block when performing 4x4 block-wise prediction coding in a method according to the present invention.

FIG. 6A is an embodiment of a table used for variable length coding a prediction method when performing 16x16 block-wise prediction coding in the method according to the present invention.

FIG. 6B is an embodiment of a table used for variable length coding a prediction method when performing 16x16 block-wise prediction coding in the method according to the present invention.

In order to achieve the above object, the image encoding method according to the present invention is the step (b) of quantizing the prediction-coded information in the step (a), step (a) after transforming by the method such as DCT And (c) entropy encoding the quantized information in step (b).

In order to achieve the above object, the apparatus for encoding an image according to the present invention converts the information encoded by the temporal / spatial predictive encoding unit for temporal / spatial predictive encoding and the temporal / spatial predictive encoding unit through a method such as DCT and the like. A transform / quantization unit for quantizing and an entropy encoding unit for entropy encoding information quantized by the transform / quantization unit are included.

In order to achieve the above object, the image decoding method according to the present invention entropy-decodes the received bit string, inversely quantizes and inverse-transforms the information obtained through entropy decoding, and the information obtained through inverse transformation. And (f) reconstructing the image through spatial / temporal prediction compensation.

In order to achieve the above object, an image decoding apparatus according to the present invention includes an entropy decoding unit for entropy decoding a received bit stream, an inverse quantization / inverse transform unit for inverse quantization and inverse transformation of information obtained through entropy decoding, and information obtained through inverse transformation. And a spatial / temporal prediction compensation unit for reconstructing the image through the spatial / temporal prediction compensation.

1. The video encoding method comprises the steps of: (a) temporal / spatial prediction coding on a single image; (b) transforming and quantizing the predictively encoded information in step (a) using a method such as DCT; (c) entropy encoding the quantized information in step (b).

2. The method of claim 1, wherein the intra image is divided into block units of a predetermined size and spatially predictive encoded using a pixel value of a neighboring block in one block.

3. The method of claim 2, wherein spatially predictive encoding is performed by referring to only pixel values of a block immediately above the current block and a block immediately to the left of the current block.

4. The method of claim 3, wherein spatial prediction is performed by referring to only pixel values of a block immediately above the current block and a block immediately to the left of the current block to simultaneously process the upper right or lower left block of the current block.

5. The method of claim 2, wherein the pixel value of the current block is predicted using the pixel value of the neighboring block according to the eight directions, but in the case of the direction referring to the pixel value of the upper right or lower left block. The value of the pixel on the right or the bottom of the block value can be used for prediction.

6. The method of item 2, wherein the encoding of the prediction method used in the current block is performed by referring to the value of the prediction method of the left or upper block.

7. The method of claim 6, wherein if there is no left or top block, or if a coding method of a different type from the current block is used, this value and reference can be regarded as using a specific one of the prediction coding methods used by the current block. The value of the prediction method used in the current block may be encoded by referring to the value of the prediction method used in another block.

8. The method of claim 6, wherein the value of the prediction method of the current block can be encoded by referring to the smaller of the values of the prediction method of the left or upper block.

9. The method of claim 8, wherein the smallest value is assigned to the most commonly used prediction method to increase the probability that the value of the prediction method used in the current block is the same as the value to which the reference value is the same.

10. The method of claim 2, when encoding the value of the prediction method used in the current block, the encoding may be performed together with other information to be encoded in the current block.

11. The method according to claim 10, wherein the probability to be used for each case of the bundled information is calculated, and according to the probability, the probable ones are encoded using short length variable lengths and the long ones using variable length codes. Can be.

2. There is provided a recording medium which records the video encoding method of items 1 to 11 as a program code executable on a computer.

13. The image encoding apparatus includes a temporal / spatial predictive encoding unit for temporal / spatial predictive encoding; A transform / quantization unit for transforming and quantizing the information predicted and encoded by the temporal / spatial predictive encoding unit through a method such as DCT; And an entropy encoder for entropy encoding information quantized by the transform / quantizer.

14. The method of claim 13, wherein the intra image may be divided into block units having a predetermined size, and spatially predictive encoding may be performed using values of pixels of neighboring blocks in one block.

15. The method of claim 14, wherein the prediction prediction coding may be performed on the basis of only pixel values of a block immediately above the current block and a block immediately to the left of the current block.

16. The method of claim 15, wherein spatial prediction encoding is performed by referring to only pixel values of a block immediately above the current block and a block immediately to the left of the current block to simultaneously process the upper right or lower left block of the current block.

17. The method of clause 14, wherein the pixel value of the current block is predicted using the pixel value of the neighboring block according to eight directions, but in the case of the direction referring to the pixel value of the upper right or lower left block. The value of the pixel on the right or the bottom of the block value can be used for prediction.

18. The method of paragraph 14, wherein the encoding of the prediction method used in the current block may be performed by referring to the value of the prediction method of the left or upper block.

19. The method of claim 18, wherein when there is no left or upper block or when a coding method of a different type is used, the value and reference may be regarded as considering one of the prediction coding methods used by the current block. The value of the prediction method used in the current block may be encoded by referring to the value of the prediction method used in another block.

20. The method of claim 18, wherein the value of the prediction method of the current block can be encoded by referring to the smaller of the values of the prediction method of the left or upper block.

21. The method of item 20, wherein the smallest value is assigned to the most commonly used prediction method to increase the probability that the value of the prediction method used in the current block is the same as the value to which the reference value is the same.

22. The method of paragraph 14, wherein when encoding a value of a prediction method used in the current block, the encoding may be performed together with other information to be encoded in the current block.

23. The method of paragraph 22, wherein the probability to be used for each case of the bundled information is calculated, and according to the probability, a long probability variable is encoded using a short length variable and a long probability variable length code. Can be.

24. An image decoding method includes: (d) entropy decoding a received bit string; (e) inverse quantization and inverse transformation of the information obtained in step (d); And (f) reconstructing the image through the spatial / temporal prediction compensation of the information obtained in step (e).

25. The method of paragraph 24, wherein the intra image is divided into block units of a predetermined size and spatially predicted and compensated using a pixel value of a neighboring block in one block.

26. The method of paragraph 25, wherein the spatial prediction is compensated by referring to only pixel values of a block immediately above the current block and a block immediately to the left of the current block.

27. The method of clause 26, wherein spatial prediction compensation can be performed by referring to only pixel values of a block immediately above the current block and a block immediately to the left of the current block, so that processing can be performed simultaneously with the upper right or lower left block of the current block. will be.

28. The method of item 27, wherein the pixel value of the current block is predicted using the pixel value of the neighboring block according to the eight directions, but in the case of the direction referring to the pixel value of the upper right or lower left block. Prediction compensation is performed using the value of the pixel on the right or the bottom of the block value.

29. The method of paragraph 25, wherein when decoding the value of the prediction method used in the current block, it may be decoded by referring to the value of the prediction method of the left or upper block.

30. The method of claim 29, wherein if there is no left or top block or if a coding method of a different type from that of the current block is used, this value and reference may be regarded as considering one of the prediction compensation methods used by the current block. It is possible to decode the value of the prediction method used in the current block by referring to the value of the prediction method used in another block.

31. The method of paragraph 29, wherein it is possible to decode the value of the prediction method of the current block by referring to the smaller of the values of the prediction method of the left or upper block.

32. The method of paragraph 31, wherein the smallest value is assigned to the most frequently used prediction method to increase the probability that the value of the prediction method used in the current block is the same as the value to which the reference value is the same.

33. The method of paragraph 25, wherein when decoding a value of the prediction method used in the current block, the decoding may be performed by combining the other information to be decoded in the current block.

34. The method of paragraph 33, wherein the probability to be used for each case of the bundled information is calculated, and the probability is deciphered by using a variable length code having a short length for a large probability and a long length for a large probability. It is preferable.

35. A recording medium is provided which records the video decoding method of paragraphs 24 to 34 in a program code executable on a computer.

36. An image decoding apparatus includes an entropy decoding unit for entropy decoding a received bit string; An inverse quantization / inverse transformation unit for inverse quantization and inverse transformation of the information decoded by the entropy decoding unit; And a spatial / temporal prediction compensation unit for restoring an image through the spatial / temporal prediction compensation from the information obtained by the inverse quantization / inverse transform unit.

37. The method of paragraph 36, wherein the intra image is divided into block units of a predetermined size, and spatial prediction prediction compensation is performed using values of pixels of neighboring blocks in one block.

38. The method of paragraph 37, wherein it is preferable to perform spatial prediction compensation by referring to only pixel values of the block immediately above the current block and the block immediately to the left.

39. The method of paragraph 38, wherein spatial prediction compensation can be performed by referring to only pixel values of a block immediately above the current block and a block immediately to the left of the current block, thereby enabling processing simultaneously with the upper right or lower left block of the current block. .

40. The method of clause 39, wherein the pixel value of the current block is predicted using the pixel value of the neighboring block according to the eight directions, but in the case of the direction referring to the pixel value of the upper right or lower left block. It is preferable to use the value of the pixel on the right or the bottom of the block value to predict and compensate.

41. The method of paragraph 37, wherein when decoding the value of the prediction method used in the current block, the decoding can be performed by referring to the value of the prediction method of the left or upper block.

42. The method of Clause 41, wherein if there is no left or top block, or if a coding method of a different type from that of the current block is used, this value and reference can be regarded as using a specific one of the prediction compensation methods used by the current block. It is preferable to decode the value of the prediction method used in the current block by referring to the value of the prediction method used in another block.

43. The method of paragraph 41, wherein it is possible to decode the value of the prediction method of the current block by referring to the smaller of the values of the prediction method of the left or upper block.

44. The method of paragraph 43, preferably assigning the smallest value to the most commonly used prediction method to increase the probability that the value of the prediction method used in the current block is the same as the value to be referred to.

45. The method of paragraph 37, wherein when decoding the value of the prediction method used in the current block, it is preferable to bind and decode together with other information to be decoded in the current block.

46. The method of paragraph 45, wherein the probability to be used for each case of the bundled information is calculated, and the probability is deciphered by using a variable length code having a short length for a large probability and a long length for a large probability. It is preferable.

Hereinafter, an image encoding and decoding method and apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

3A is a block diagram schematically illustrating an embodiment of an image encoding apparatus according to the present invention, and includes a temporal / spatial prediction encoding unit 100, a transform / quantization unit 101, and an entropy encoding unit 102. It is composed.

3B is a block diagram schematically illustrating an embodiment of an image decoding apparatus according to the present invention, which includes an entropy decoder 110, an inverse quantization / inverse transform unit 111, and a spatial / temporal prediction compensation unit 112. It is composed.

4A is a flowchart illustrating an example of an image encoding method performed by the apparatus of FIG. 3A, and FIG. 4B is a flowchart illustrating an example of an image decoding method performed by the apparatus of FIG. 3B.

Referring to FIGS. 3A and 4A, the temporal / spatial prediction encoding unit 100 performs spatial prediction encoding on an intra image (step 200) and performs temporal prediction encoding on an inter image (step 201). . In the case of spatial predictive coding, the luminance component and the color component are predictively encoded, and the prediction is performed by using the pixel values of the neighboring blocks. The transform / quantization unit 101 loss-compresses the information by transforming and quantizing the predictive coded information through a transform such as DCT (step 202). The entropy encoder 102 generates a bit string by entropy encoding the quantized information (step 203).

3B and 4B, the entropy decoding unit 110 entropy decodes the received bit string (operation 210). The inverse quantization / inverse transform unit 111 performs inverse quantization and inverse transformation on the entropy-decoded information (step 211), and reconstructs the image by performing prediction compensation on the spatial / temporal prediction compensation unit 112 again (step 212). step).

5 is a view for explaining a spatial prediction method according to the present invention. In the JVT final committee's recommendation, prediction was made using 17 pixel values of ABCD EFGH IJKL MNOP Q with reference to FIG. 1B. Conduct nine methods of forecasting, such as the final committee recommendations. However, among the nine methods, three methods using EFGH MNOP were modified as follows.

Method 3: diagonal down / left prediction

a is (A + 2B + C + I + 2J + K + 4) >> 3

b, e is (B + 2C + D + J + 2K + L + 4) >> 3

c, f, i are (C + 3D + K + 3L + 4) >> 3

d, g, j, m, h, k, n, l, o, p are predicted using (D + L + 1) >> 1.

Method 7: vertical-left prediction

a is (2A + 2B + J + 2K + L + 4) >> 3

b, i is (B + C + 1) >> 1

c, j is (C + D + 1) >> 1

e is (A + 2B + C + K + 3L + 4) >> 3

f, m is (B + 2C + D + 2) >> 2

g, n is (C + 3D + 2) >> 2

d, k, l, h, o, p are predicted using D.

Method 8: horizontal-up prediction

a is (B + 2C + D + 2I + 2J + 4) >> 3

b is (C + 3D + I + 2J + K + 4) >> 3

c, e is (J + K + 1) >> 1

d, f is (J + 2K + L + 2) >> 2

g, i is (K + L + 1) >> 1

h, j is (K + 3L + 2) >> 2

l, n, k, m, o, p are predicted using L.

The advantage of using only nine pixels as described above is that the current block and the upper right and lower left blocks can be processed simultaneously because there is no need to refer to the values of the upper right and lower left blocks of the current block. In addition, compared to the existing method, the amount of calculation is reduced when the above modification is made.

When performing the 4x4 block spatial prediction and encoding the prediction method into the bit string, the JVT final committee recommendation, if the above block or the left block is not encoded by the intra method or cannot be referred to, the value to be referred to is DC. In the present invention, if the one of the above blocks or the left block is not encoded by the intra method or cannot be referred to, the block is regarded as the DC method and the above block or the left block is set in the present invention. Compared to the prediction method used in the block that can be referred to, it is set to a value referring to the smaller of the two.

This method is particularly effective at the boundary of an image or slice. For example, if the above block cannot be referred to when encoding with the intra method at the upper boundary of an image or slice, it is referred to to encode the prediction method of the current block. The value is the DC method according to the JVT final committee recommendations. However, in the method according to the present invention, the DC method is compared with the value of the prediction method used in the block on the left side of the current block to refer to a smaller value. If the left block and the current block used the same horizontal prediction method, the JVT final committee recommendation method did not use the value referred to using 1 bit to encode the prediction method used in the current block, that is, the DC prediction method. After coding, we have to code that the horizontal prediction method is used using 3 bits. However, in the method according to the present invention, the value referred to is not set by the DC method unconditionally, but is smaller than the value of the left block, that is, the value referred to by the horizontal method. You can tell.

In addition, in the JVT final committee recommendation, the vertical method has the smallest value of 0 and the horizontal method has the smallest value of 1. In general, however, the horizontal method is used more frequently than the vertical method is used. This statistic is especially pronounced in the case of interlaced images. This is because the interlace image is divided into odd and even fields so that the similarity is higher in the horizontal direction than in the vertical direction.

Based on these statistics, it is preferable to set the horizontal method to 0 and the vertical method to 1 to increase the efficiency of the method of encoding which prediction method is used in units of 4x4 blocks. In the method according to the present invention, the horizontal method is set to 0 and the vertical method is set to 1 to improve the compression efficiency. The number of other encoding methods can also be set by such a statistics method.

In addition, (Imode, nc, AC) is encoded to encode which prediction method is used in units of 16x16 blocks. According to statistical data, the probability that nc is 2, that is, both DC and AC in color components is very small. . Experiments show that the probability that nc is 0 or 1 is close to 90%. However, since (Imode, nc, AC) uses a variable length code, it is preferable to use a longer code when the probability is small. The JVT final committee recommendation prevents the assignment of variable-length codes by this probability because nc is centered in (Imode, nc, AC). Therefore, the method according to the present invention improves the efficiency of variable length coding by encoding in the order of (Imode, AC, nc). The variable length code table thus changed is shown in FIG. 6A. Alternatively, (Imode, nc, AC) may be encoded, and when nc is 2, the variable length codes may be changed in order as shown in FIG. 6B so that a longer variable length code is used.

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 disks, optical data storage devices, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet). Include. 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.

The best embodiments have been disclosed in the drawings and specification above. Although specific terms have been used herein, they are used only for the purpose of describing the present invention and are not used to limit the scope of the present invention as defined in the meaning or claims. Therefore, those skilled in the art will understand that various modifications and equivalent other embodiments are possible from this. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

As described above, according to the video encoding and decoding method and apparatus according to the present invention, by limiting the pixels used for prediction in 4x4 spatial prediction encoding to the upper block and the left block of the current block, the calculation amount is reduced and the upper right or the left Simultaneous processing with the bottom block was made possible. In addition, when encoding which prediction method is used in units of 4x4 blocks or 16x16 blocks, compression efficiency is increased by effectively encoding in consideration of statistical characteristics.

Claims (1)

(a) temporal / spatial prediction coding on a single image; (b) transforming and quantizing the predictively encoded information in step (a) using a method such as DCT; (c) entropy encoding the quantized information in step (b).