RU2314656C2 - Internal encoding, based on spatial prediction - Google Patents

Abstract

FIELD: encoding of inventions and, in particular, encoding of video frame blocks.

SUBSTANCE: method and device are suggested, meant for encoding digital image with usage of prediction block in internal mode, where the list of prediction modes is produced for each combination of prediction modes of adjacent blocks. Modes, meant for each combination of prediction modes, may be divided on two groups. First group has m most probable prediction modes, and second group has (n-m) prediction modes, where n is common number of available prediction modes. Modes in the first group are ordered in accordance to their probability. Aforementioned order may be set as a list of modes, ordered from most probable one to least probable one. Modes, belonging to the second group, may be ordered using a certain given method, which may be set depending on information which is already available in the decoder.

EFFECT: reduced memory volume with minimal loss of encoding efficiency.

6 cl, 6 tbl, 8 dwg

Description

Description

FIELD OF THE INVENTION

The present invention generally relates to image encoding, and more particularly to encoding blocks of video frames.

State of the art

A digital image, such as a video image, a television image, a still image or an image generated by a VCR or computer, consists of pixels arranged in horizontal and vertical lines. The number of pixels in a single image is usually tens of thousands. Each pixel typically contains luminance and color information. Without compression, the amount of information transmitted from the image encoder to the image decoder is so huge that it makes real-time image transmission impossible. To reduce the amount of information transmitted, a number of different compression methods have been developed, such as JPEG, MPEG and H.263 standards. In a typical video encoder, the frame of the original video sequence is divided into rectangular areas or blocks that encode in internal mode (mode I) or in intermediate mode (mode P). Blocks are independently encoded using some form of transform coding, such as DCT encoding, DCT (Discrete Cosine Transform). However, simple block-based coding only reduces the correlation between pixels in a particular block without taking into account inter-block correlation of pixels and it still creates high bit rates for transmission. Current digital image coding standards also use certain methods that reduce the correlation of pixel values between blocks.

Typically, blocks encoded in P mode are predicted from one of the previously encoded and transmitted frames. Block prediction information is represented using a two-dimensional (2D) motion vector. For blocks encoded in mode I, the predicted block is formed using spatial prediction from already encoded neighboring blocks in the same frame. A prediction error, for example, the difference between a coded block and a predicted block, is represented as a set of weighted basic functions of some discrete transform. The conversion is usually performed on the basis of an 8x4 or 4x4 block. Weights — Transformation coefficients are then quantized. Quantization introduces a loss of information and, therefore, quantized coefficients have lower accuracy than the originals.

Quantized transform coefficients, together with motion vectors and some control information, form a complete representation of the encoded sequence and are called syntactic elements. Before transferring from the encoder to the decoder, all syntax elements are encoded with respect to entropy in such a way as to further reduce the number of bits required to represent them.

In the decoder, the block in the current frame is obtained first by compiling its prediction in the same way as in the encoder, and by adding a compressed prediction error to the prediction. A compressed forecast error is found by weighting the basic transformation functions using quantized coefficients. The difference between the recovered frame and the original frame is called a recovery error.

Compression ratio, i.e. the ratio of the number of bits used to represent the original and compressed sequences, both in the case of blocks I and P, is controlled by adjusting the magnitude of the quantization parameter, which is used to quantize the transform coefficients. The compression ratio also depends on the entropy encoding method used.

An example of spatial prediction used in the working draft number 2 (WD2, RP2) of the JVT encoder is described as follows. In order to perform spatial prediction, the IVT encoder offers 9 modes for predicting 4x4 blocks, including DC prediction (mode 0) and 8-direction modes, marked 1 through 7, as shown in FIG. 1. The forecasting process is illustrated in figure 2. As shown in FIG. 2, pixels from a to p must be encoded, and pixels from A to Q from neighboring blocks that are already encoded are used for prediction. If, for example, mode 1 is selected, then pixels a, e, i and m are predicted by setting them equal to pixel A, and pixels b, f, j and n are predicted by setting them equal to pixel B, etc. Similarly, if mode 2 is selected, pixels a, b, c and d are predicted by setting them equal to pixel I, and pixels e, f, g and h are predicted by setting them equal to pixel J, etc. Therefore, mode 1 is predictive mode in the vertical direction, and mode 2 is a predictive mode in the horizontal direction. These modes are described in document VCEG-N54 published by the telecommunications standardization sector of the ITU video coding expert group (VCEG) in September 2001, and in document JVT-B118r2 published by the joint ISO / IEC MPEG video group and ITU-T VCEG in March 2002

Mode 0: DC Prediction

Typically, all samples are predicted using (A + B + C + D + I + J + K + L + 4) >> 3. If four of the samples are outside the image, the average of the other four is used for forecasting. If all eight samples are outside the image, the prediction for all samples in the block is 128. Therefore, the block can always be predicted in this mode.

Mode 1: Vertical Prediction

If A, B, C, D are inside the image, then

- a, e, i, m are predicted using A,

- b, f, j, n are predicted using B,

- c, g, k, o are predicted using C,

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

Mode 2: Horizontal Prediction

If E, F, G, H are inside the image, then

- a, b, c, d are predicted using E,

- e, f, g, h are predicted using F,

- i, j, k, l are predicted using G,

- m, n, o, p are predicted using H.

Mode 3: Diagonal Prediction Down / Right

This mode is used only if all A, B, C, D, I, J, K, L, Q are inside the image. This is a "diagonal" prediction.

- m is predicted using (J + 2K + L + 2) >> 2

- i, n are predicted using (I + 2J + K + 2) >> 2

- e, j, o are predicted using (Q + 2I + J + 2) >> 2

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

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

- c, h are predicted using (A + 2B + D + 2) >> 2

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

Mode 4: Diagonal Prediction Down / Left

This mode is used only if all A, B, C, D, I, J, K, L, Q are inside the image. This is a "diagonal" prediction.

- and they predict using (A + 2B + C + I + 2J + K + 4) >> 3

- b, e are predicted using (A + 2B + C + I + 2J + K + 4) >> 3

- c, f, i are predicted using (C + 2D + E + R + 2L + M + 4) >> 3

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

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

- l, o are predicted using (F + 2G + H + N + 2O + P + 4) >> 3

- p is predicted using (G + H + O + P + 2) >> 3

Mode 5: Vertical-Left Prediction

This mode is used only if all A, B, C, D, I, J, K, L, Q are inside the image. This is a "diagonal" prediction.

- a, j are predicted using (Q + A + 1) >> 1

- b, k are predicted using (A + B + 1) >> 1

- c, l are predicted using (B + C + 1) >> 1

- d is predicted using (C + D + 1) >> 1

- e, n are predicted using (I + 2Q + A + 2) >> 2

- f, o are predicted using (Q + 2A + B + 2) >> 2

- g, p are predicted using (A + 2B + C + 2) >> 2

- h is predicted using (B + 2C + D + 2) >> 2

- i predict using (Q + 2I + J + 2) >> 2

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

Mode 6: vertical-right prediction

This mode is used only if all A, B, C, D, I, J, K, L, Q are inside the image. This is a "diagonal" prediction.

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

- b, i are predicted using (B + C + 1) >> 1

- c, j are predicted using (C + D + 1) >> 1

- d, k are predicted using (D + E + 1) >> 1

- l is predicted using (E + F + 1) >> 1

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

- f, m are predicted using (B + 2C + D + 2) >> 2

- g, n are predicted using (C + 2D + E + 2) >> 2

- h, o are predicted using (D + 2E + F + 2) >> 2

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

Mode 7: horizontal upward prediction

This mode is used only if all A, B, C, D, I, J, K, L, Q are inside the image. This is a "diagonal" prediction.

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

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

- c, e are predicted using (D + 2E + F + 2J + 2K + 4) >> 3

- d, f are predicted using (E + 2F + G + J + 2K + 2L + 4) >> 3

- g, i are predicted using (F + 2G + H + 2K + 2L + 4) >> 3

- h, j are predicted using (G + 3H ++ K + 3L + 4) >> 3

- l, n are predicted using (L + 2M + N + 2) >> 3

- k, m are predicted using (G + H + L + M + 2) >> 2

- o predict using (M + N + 1) >> 1

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

Mode 8: Horizontal Prediction Down

This mode is used only if all A, B, C, D, I, J, K, L, Q are inside the image. This is a "diagonal" prediction.

- a, g are predicted using (Q + I + 1) >> 1

- b, h are predicted using (I + 2Q + A + 2) >> 2

- c is predicted using (Q + 2A + B + 2) >> 2

- d is predicted using (A + 2B + C + 2) >> 2

- e, k are predicted using (I + J + 1) >> 1

- f, l are predicted using (Q + 2I + J + 2) >> 2

- i, o are predicted using (J + K + 1) >> 1

- j, p are predicted using (I + 2J + K + 2) >> 2

- m is predicted using (K + L + 1) >> 1

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

Since each block must have a prediction mode assigned and transmitted to the decoder, this would require a significant number of bits if encoded directly. In order to reduce the amount of information transmitted, correlation of prediction modes of neighboring blocks can be used. For example, a patent application owned by Vateri et al. (WO 01/54416 A1 “Image encoding method and image encoder”, hereinafter referred to as Vateri), discloses block-based coding method that uses image directivity information in blocks to classify many spatial prediction modes. This spatial prediction mode of the block is determined using the directivity information of at least the neighboring block.

In the JVT encoder, when the prediction modes of neighboring already encoded blocks U and L are known, the most probable prediction mode, the next most probable prediction mode, etc. are ordered for block C (Fig. 3). The ordering of the modes is set for each combination of the prediction modes U and L. This ordering can be defined as a list of prediction modes for block C, ordered from the most probable to the least probable mode. The ordered list used in RP2 of the JVT encoder, as disclosed in VCEG-N54, is shown below:

Table I Prediction mode as a function of ordering, transmitted as a signal in a bit stream L / U Beyond 0 one 2 3 Beyond -------- 0 -------- 01 ------- 10------- --------- 0 02 ------- 021648573 125630487 021876543 021358647 one --------- 102654387 162530487 120657483 102536487 2 twenty------- 280174365 217683504 287106435 281035764 3 --------- 201385476 125368470 208137546 325814670 four --------- 201467835 162045873 204178635 420615837 5 --------- 015263847 152638407 201584673 531286407 6 --------- 016247583 160245738 206147853 160245837 7 --------- 270148635 217608543 278105463 270154863 8 --------- 280173456 127834560 287104365 283510764 L / U four 5 6 7 8 Beyond -------- --------- --------- --------- --------- 0 206147583 512368047 162054378 204761853 208134657 one 162045378 156320487 165423078 612047583 120685734 2 287640153 215368740 216748530 278016435 287103654 3 421068357 531268470 216584307 240831765 832510476 four 426015783 162458037 641205783 427061853 204851763 5 125063478 513620847 165230487 210856743 210853647 6 640127538 165204378 614027538 264170583 216084573 7 274601853 271650834 274615083 274086153 278406153 8 287461350 251368407 216847350 287410365 283074165

In the present description, an example of prediction modes for block C is given as defined in RP2 of the JVT encoder when the prediction mode for both U and L is 2. The sequence (2,8,7,1,0,6,4,3, 5) indicates that mode 2 is also the most likely mode for block C. Mode 8 is the next most likely mode, etc. Information will be transmitted to the decoder indicating that the nth most probable mode will be used for block C. The ordering of the modes for block C can also be specified by listing the rank for each mode: the higher the rank, the less likely the forecasting method. For the above example, the list of ranks would be (5,4,1,8,7,9,6,3,2). When the modes (0,1,2,3,4,5,6,6,7,8) are associated with the list of ranks (5,4,1,8,7,9,6,3,2), we can say that the mode 0 has rank 5, mode 1 has rank 4, etc.

For more efficient coding, intra prediction information of two 4x4 blocks may be encoded in one codeword.

The aforementioned method has one main drawback - the memory required to store the ordering of prediction modes for block C is required, provided that the prediction modes of blocks U and L are present. and L. For each combination, the ordering of 9 possible modes must be specified. This means that 9x9x9 bytes are required (in the present description, the applicant assumes that one number requires one byte) to specify the ordering of the prediction modes. In addition, more memory may be required to specify special cases, for example, if one or both of the U and L blocks are unavailable.

Therefore, it is advantageous and desirable to provide a method and apparatus for encoding a digital image in which memory requirements are reduced, while the loss in encoding efficiency is minimal.

SUMMARY OF THE INVENTION

A method and apparatus are proposed for encoding a digital image using block prediction of the internal mode. Get a list of prediction modes for each combination of prediction modes of neighboring blocks. The modes assigned to each combination of prediction modes can be divided into two groups. The first group includes m (where m is less than the total number n of available modes) of the most probable forecasting modes, and the second group includes the remaining modes. The modes in the first group are ordered according to their probability. This ordering can be defined as a list of modes ordered from most probable to least probable. The modes belonging to the second group can be ordered in some specified way, which can be set depending on the information already available in the decoder. Information regarding whether the mode selected for a given block belongs to the first group or to the second group is sent to the decoder. If it belongs to the first group, information is transmitted indicating that the i-th most probable mode will be used for block C under the condition of this combination of modes for blocks U and L. If the mode belongs to the second group, information is transmitted indicating what should be used j-th mode of this group.

The present invention will become apparent after reading the description taken in conjunction with FIG. 4a to 8.

Brief Description of the Drawings

Figure 1 is a schematic diagram illustrating 8 directional modes that are used as spatial prediction modes.

2 is a schematic diagram illustrating pixels that are used to predict the current block of 4x4 pixels.

3 is a schematic diagram illustrating two neighboring blocks used to predict the current block.

4a is a schematic diagram illustrating a spatial prediction mode of two neighboring blocks used to predict the current block.

Fig. 4b is a schematic diagram illustrating a spatial prediction mode of two neighboring blocks having a mirror image with the blocks of Fig. 4a.

Fig. 5a is a schematic diagram illustrating another spatial prediction mode.

5b is a schematic diagram illustrating a mirrored mode pair.

6 is a flowchart illustrating a spatial prediction method.

7 is a block diagram illustrating a transmission system of a digital image block for implementing a method in accordance with an embodiment of the present invention.

FIG. 8 is a flowchart illustrating a portable video telecommunications device implementing the method in accordance with an embodiment of the present invention.

The best way to carry out the invention

An embodiment of the present invention uses the property that it is possible to obtain an ordered list of prediction modes for one combination of prediction modes of neighboring blocks as a function of prediction modes of another combination. To illustrate the prediction modes of two neighboring blocks U and L, as shown in Fig. 3, they are used to infer the prediction of the current block C. It should be noted that the combination of prediction modes in Fig. 4a can be obtained by inverting the prediction modes relative to the diagonal, as shown in fig.4b. Thus, the nth most probable prediction mode for block C when the combination of modes in FIG. 4b. Therefore, if the neighboring blocks U and L have the “vertical” and “vertical” modes, the prediction mode of the current block is most likely “vertical” (Fig. 4b). Therefore, when these blocks are “facing” or mirrored relative to the diagonal (“down / right”), it is known that “horizontal” for the current block should be obtained from “horizontal” and “horizontal” (Fig. 4a). Similarly, if adjacent blocks U and L are mode blocks 2 and 3, as shown in FIG. 5a, then the inverted blocks U and L will be blocks of modes 3 and 1, as shown in FIG. 5b.

In order to further illustrate this example, the applicant defines a function f that maps the prediction direction i to j, j = f (i). Each prediction mode i is assigned a prediction mode j obtained by mirroring it relative to a diagonal line extending from the upper left corner of the block to the lower right corner of the block. For the prediction modes in FIG. 1 the resulting assignment is summarized in table II.

Table II i j 0 0 one 2 2 one 3 3 four four 5 8 6 7 7 6 8 5

When a function is defined as described above, an ordered list of prediction modes for a combination of modes (k, l) can be determined based on an ordered list for a combination of (i, j) such that i = f (l) and j = f (k ), i.e., if the prediction mode p is the nth most probable mode for the combination (i, j), the nth mode for the combination (k, l) is f (p). As an example, the applicant considers a combination of modes (1,1), which is assigned an ordered list of modes for block C: (1,6,2,5,3,0,4,8,7). It would be possible to obtain an ordered list of prediction modes for a combination of (2.2) from this ordered list using a mapping using the function f: (2,7,1,8,3,0,4,6,5). Similarly, the ordered list of prediction modes for combination (2,3) is (2,0,8,1,3,7,5,4,6), and the ordered list of modes f (2,3) = (3,1) is f (2,0,8,1,3,7,5,4,6) = (1,0,5,2,3,6,8,4,7). It should be noted that the ordered list of prediction modes for (k, l) can essentially be symmetric to the list for (i, j). Therefore, the display function f can be described as a mirror image function.

The primary objective of the present invention is to reduce the size of a table defining a prediction mode as a function of prediction modes of neighboring blocks already encoded (e.g., table 1). This table is obtained during the learning process. To get an ordered list of prediction modes for a combination of modes (i, j) for a large number of blocks C for which neighboring blocks have this combination, count the number when each prediction mode is selected in accordance with this selection criterion (for example, the smallest prediction error) . This number determines the rank of the prediction mode in the list of prediction modes assigned to the combination (i, j), the more often the mode is selected, the higher the rank that it has in the list of prediction modes.

When the described mapping of prediction modes is used during the learning process, the resulting table defining the prediction modes will be smaller. If the ordered list for the combination (k, l) can be obtained from the ordered list for the combination (i, l), only the ordered list for the combination (i, j) should be remembered. During training, the display can be represented as follows. The frequency of occurrence of each prediction mode for combinations (k, l) and (i, j), such that i = f (l) and j = f (k), are counted together, i.e., if the prediction mode p was chosen for combination (k, l), it is also believed that the prediction mode f (p) was selected for combination (i, j). Also, when the prediction mode p was selected for the combination (i, j), it is believed that the prediction mode f (s) was selected for the combination (k, l).

A reduced table in accordance with the present invention is shown in table III.

Table III Reduced Prediction Table L / U Beyond 0 one 2 3 Beyond -------- 0 -------- 01 ------- 10------- --------- 0 02 ------- 024167835 150642387 027486135 013245867 one --------- 150643278 021468735 105436287 2 twenty------- 124086573 283407156 3 --------- 385240167 four --------- 5 --------- 6 --------- 7 --------- 8 --------- L / U four 5 6 7 8 Beyond  -------- --------- --------- --------- --------- 0  012465738  150346287  160452387  024716835 028413765 one  104562378  156403287  165403278  014652738 014256837 2  240781635  214835076  241086735  207483165 280473156 3  413205876  531480267  146530287  247308516 832045176 four  420671835  145602387  461027538  407261835 248073165 5  513406287  165402387  240158376 082354167 6  614503287  614057328  042617385 024681357 7  427016385  426701835 284703615 8 328514067  248361075  248703651

Table III does not give an ordered list of prediction modes for some combinations (U, L). Ordered lists for these combinations can be "restored" by displaying the corresponding elements that are left in the prediction table when these "restored" elements are needed to predict the current block. Therefore, usually, since an element in the prediction table can be obtained or retrieved from another element in the prediction table by display, the first element can be excluded. In other words, in the prediction table containing the first group of elements and the second group of elements, each element of the second group of elements can be restored from the corresponding element of the first group using the display function, the second group of elements can be excluded.

6 is a flowchart illustrating a decoding step when symmetry is used in the prediction table. As depicted, the method 100 comprises a step 110 in which a plurality of image blocks are received. When the current block is processed, it is determined in step 120 whether the prediction mode of the current block can be obtained from the prediction mode of neighboring blocks without displaying. If this is true, then the spatial prediction mode of the current block is determined based on the prediction mode of neighboring blocks in step 132. Otherwise, an additional prediction mode of neighboring blocks is provided in step 130, and an additional prediction mode of the current block is determined based on the additional prediction mode of neighboring blocks in step 140. At step 150, an additional prediction mode of the current block is displayed in the prediction mode of the current block.

Alternatively, it is possible to assign the same label to different prediction modes (grouping them together) of the U and L blocks before using them to specify the prediction mode for block C. For example, in the case of the JVT encoder, modes 1,5 and 6 can be grouped together and marked as 1, and modes 2,7 and 8, can be grouped together and marked as 2. As can be seen from figure 1, the directions of modes 7 and 8 are close to the direction of mode 2, and the directions of modes 5 and 6 are close to mode direction 1. After this grouping, each of the blocks U and L can have one of 5 modes marked as 0,1,2,3 and 4. Therefore, instead of 9x9 possible combinations of prediction modes U and L, there are only 5x5 of such combinations. Thus, the memory required to specify the ordering of prediction modes for block C under the given data of prediction modes of blocks U and L will be 5x5x9 bytes instead of 9x9x9 bytes (assuming that 1 byte of memory is needed in order to store 1 digit). In addition, if the display function f is used to “reverse” the ordered list, the prediction table can be further simplified.

An example of a table defining a prediction mode as an ordering function transmitted as a signal in a bit stream when both of these methods are used together is presented in Table IV.

Table IV L / U Beyond 0 one 2 3 four Beyond -------- 0 -------- 01 ------- 10------- --------- --------- 0 02 ------- 024167835 150642387 024781635 013245867 012465738 one --------- 156043278 021468375 153046827 140652378 2 twenty------- 214806357 283407156 247081635 3 --------- 385240167 413205876 four --------- 420671835

In addition, it is also possible to limit the number of prediction modes for block C provided that these prediction modes of blocks U and L. In the case of the JVT encoder, there would still be 9x9 possible combinations of prediction modes U and L. But only m would be assigned to each of these combinations modes, where m is less than 9. Thus, the number of possible prediction modes is reduced to (9x9xm) <(9x9x9). Similarly, if the mapping function f is used to “reverse” ordered lists, the prediction table can be further simplified.

These methods can be used together or separately.

Spatial prediction-based intra-coding in accordance with the present invention can be easily integrated into a digital image block transmission system, as shown in FIG. Assuming that the frame must be encoded in an internal format using some kind of internal prediction, the frame is encoded as follows. The blocks of the encoded frame are sent sequentially to the encoder 50 of the video transmission system shown in Fig.7. The frame units are received from a digital image source, such as a camera or VCR (not shown), at the input 27 of the image transmission system. In a method known per se, blocks received from a digital image source comprise pixel values of an image. A frame may be temporarily stored in a frame memory (not shown) or, alternatively, the encoder receives input directly block by block.

The blocks are sequentially sent to the prediction method selection block 35, which determines whether the pixel values of the current block to be encoded can be predicted based on blocks previously internally encoded in the same frame or segment. In order to do this, the prediction method selection unit 35 receives input from the frame buffer of the encoder 33, which contains a record of previously encoded and then decoded and reconstructed indoor units. Thus, the prediction method selection unit may determine whether prediction of the current block can be performed based on previously decoded and reconstructed blocks. In addition, if there are corresponding decoded blocks, the prediction method selection unit 35 may select the most suitable method for predicting the pixel values of the current block if more than one such method can be selected. It should be understood that in certain cases, predicting the current block is impossible, since suitable blocks for use in prediction are not in the frame buffer 33. In a situation where there is more than one prediction method, information about the selected prediction method is supplied to the multiplexer 13 for further transmission to the decoder. It should also be noted that in some prediction methods, certain parameters necessary to perform prediction are transmitted to the decoder. Of course, this depends on the exact implementation adopted and in no way limits the prediction-based intra-coding application of the present invention.

The pixel values of the current block are predicted in the intra prediction block 34. The intra prediction unit 34 receives input related to the selected prediction method from the prediction method selection block 35 and information related to the blocks available for use in prediction from the frame buffer 33. Based on this information, the intra prediction block 34 makes a forecast for the current block. The predicted pixel values for the current block are sent to a differential adder 28, which creates a prediction error block by taking the difference between the pixel values of the predicted current block and the actual pixel values of the current block received from input 27. Then, the error information for the predicted block is encoded in the encoding block forecasting errors in an effective form for transmission, for example, using a discrete cosine transform (DCT). The encoded prediction error block is sent to the multiplexer 13 for further transmission to the decoder. The encoder of the digital image transmission system also includes decoding functionality. The encoded prediction error of the current block is decoded in the decoding block 30 of the prediction errors, and then summed in the adder 31 with the predicted pixel values for the current block. Thus, an encoded version of the current block is obtained. The encoded current block is then sent to a frame buffer 33.

It is also contemplated herein that a receiver receives blocks that form a digital image frame on only one of the transmission channels.

At receiver 60, the demultiplexer receives blocks of the demultiplexed encoded prediction error and prediction information transmitted from the encoder 50. Depending on the prediction method in question, the prediction information may include parameters used in the prediction process. It should be understood that in the case where only one prediction method is used, information related to the prediction method used to encode blocks is optional, although it may still be necessary to transmit parameters used in the prediction process. 7, dashed lines are used to represent optional transmission and reception of information about the prediction method and / or prediction parameters. Assuming that more than one prediction method can be used, information regarding the selection of the prediction method for the current decoded block is supplied to the intra prediction block 41. The intra prediction unit 41 examines the contents of the frame buffer 39 to determine if previously encoded blocks exist for use in predicting the pixel values of the current block. If such image blocks exist, the intra prediction block 41 predicts the contents of the current block using the prediction method indicated by the received prediction method information, and possibly using the prediction related parameters received from the encoder. Prediction error information associated with the current block is received using the prediction error decoding unit 36, which decodes the prediction error block using a suitable method. For example, if the prediction error information was decoded using a discrete cosine transform, the prediction error decoding unit performs the inverse DCT to extract the error information. The prediction error information is then summed with the prediction for the current image block in the adder 37 and the output of the adder is supplied to the frame buffer 39. In addition, since each block is decoded, it is sent to the output of the decoder 40, for example, for display on some form of display means. Alternatively, an image frame can only be displayed after the entire frame is decoded and stored in frame buffer 39.

It should be noted that the intra prediction unit 34 makes a prediction of the current block based on previously encoded and then decoded and reconstructed indoor blocks, as provided by the frame buffer 33. In particular, the prediction of the current block is determined from the spatial prediction modes of previously reconstructed indoor blocks using the prediction table, as shown in table III or table IV (not shown in Fig. 7). However, when the ordered list for the prediction modes (i, j) of previously reconstructed indoor units is omitted in the prediction table, the conversion unit 32 can be used to display the spatial prediction modes of the previously reconstructed blocks into additional or mirrored spatial prediction modes (k, l). At this point, the intra prediction unit 34 may determine an additional or mirrored mode f (p) for the current block. Again, the display unit 32 is used to obtain the prediction mode p of the current block by displaying the additional prediction mode f (p). In the same way, the display unit 38 is used for display when necessary.

The mapping algorithm that is used to map (i, j) to (k, l) and map f (p) to p can be encoded in software program 69, which contains machine-accessible steps or pseudo-codes for executing method in accordance with the present invention. Advantageously, the software program is stored on a storage medium. For example, the software program is stored in a memory device located in a CPU, a CPU 70, or in a separate memory device 68, as shown in FIG. FIG. 8 is a simplified schematic diagram of a mobile terminal 90 for use as a portable video telecommunications device implementing the prediction mode display method of the present invention. The mobile terminal 90 includes at least a display module 76 for displaying images, an image pickup device 72 and an audio module 74 for capturing audio information from an audio input device 82 and reproducing audio information on an audio playback device 80. Advantageously, the movable terminal 90 further comprises a keyboard 70 for inputting data and commands, a radio frequency component 64 for communicating with the mobile telecommunications network, and a signal / data processing device 70 for controlling the operation of the telecommunication device. Preferably, the transmission system (50, 60) of the digital image block is implemented in the processor 70.

In accordance with a further embodiment of the present invention, memory requirements could be limited in obtaining coding efficiency.

The applicant noted that for each combination of modes for blocks U and L there are only a few modes for block C, which have a high probability of occurrence. The likelihood of other modes is much less. In addition, the probabilities of the appearance of these other modes are the same and, therefore, their ordering in an ordered set of modes does not have a strong influence on the coding efficiency.

The modes assigned to each combination of prediction modes U and L can be divided into two groups. The first group includes m (where m is less than the total number n of available modes) of the most probable forecasting modes, and the second group includes the remaining modes. The modes in the first group are ordered according to their probability. As described above, this ordering can be defined as a list of modes ordered from most probable to least probable. These modes, belonging to the second group, can be ordered in some specified way, which can be set depending on the information already available for the decoder. For example, a prediction mode that is marked using a smaller number may precede a mode that is marked using a larger number.

In other words, prediction modes are located in an ordered set S. An ordered set of R modes assigned to combinations of prediction modes for U and L is created by combining two ordered sets of modes T = {t1, t2, ..., tk} and V = { v1, v2, ..., vl}:

R = {t1, t2, ..., tk, v1, v2, ..., vl}.

The modes in the first ordered set T are ordered according to their expected probability of occurrence for the combination of modes of neighboring blocks U and L. The more often the mode is expected, the less bits should be used to transmit a signal about it. The second ordered set of modes V is created by removing the elements of the first set T from the set S and maintaining the ordering of the remaining elements. For example, if prediction i precedes prediction mode j in ordered set S, prediction i must precede prediction mode j in set V.

As a separate example, the applicant considers the modes available in the JVT encoder and assumes that for each combination of modes for blocks U and L, only the most probable mode is assigned to the first group. The remaining 8 modes are listed in increasing order of their numbers. To further illustrate this separate example, the applicant considers a combination of modes (1,1), which is assigned to mode 1 as the most likely mode. Mode 1 is the first ordered set S, that is, V = {0,2,3,4,5,6,7,8}. In this example, an ordered set S is created by listing nine prediction modes in an increasing sequence of numbers, that is, S = {0,1,2,3,4,5,6,7,8}.

Information regarding whether the mode selected for a given block belongs to the first group or to the second group is sent to the decoder. If it belongs to the first group, information is transmitted indicating that the i-th most probable mode will be used for block C, provided this combination of modes for blocks U and L. In this separate example, if the decoder receives information that the first group, no additional information is required, since only one mode belongs to this group. If the mode belongs to the second group, information is transmitted indicating that the j-th mode of this group should be used. In this separate example, if information is received that the first mode should be selected, mode 0 should be selected.

Alternative embodiments:

1. The rules according to which the modes are ordered in the set S may differ.

2. The method by which the elements of the first set T are determined depends on the prediction modes of the neighboring blocks U and L. Elements of this set can be specified, for example, as a table stored in the memory of both the encoder and the decoder, or they can be logically deduced from the modes U and L.

3. The transmission of the signal to the decoder regarding which mode is selected can be performed in numerous ways, for example:

- information about the rank of the prediction mode in the ordered set R can be transmitted;

- for information regarding which set - T or V is selected, the rank of the mode in the selected set follows.

A preferred implementation for the prediction modes specified in the JVT encoder is described as follows.

The same label can be assigned to different prediction modes of blocks U and L before using them to specify the prediction mode for block C. The diagonal modes 3, 5 and 8 are grouped together and marked as 3, and the diagonal modes 4, 6 and 7 are grouped together and are designated as 4. After this grouping, each of the blocks U and L can have one of five modes, designated as 0, 1, 2, 3, and 4. Therefore, instead of 9x9 possible combinations of prediction modes U and L, there are only 5x5 of such combinations.

An ordered set S is created by listing nine prediction modes in an increasing sequence of their numbers, that is, S = {0,1,2,3,4,5,6,7,8}. For each combination of prediction modes for U and L, only one most probable mode is assigned, i.e., the ordered set T for each combination consists of only one element. The assignment is given in table V.

Table v L / U Beyond 0 one 2 3 four Beyond 0 0 one 0 0 0 one 0 0 0 one 0 one 0 one one 2 2 2 one 2 2 2 3 0 one 2 3 four four 0 one 2 four four

The first information regarding whether the most probable mode is used is sent to the decoder. This information is coded together for two 4x4 blocks, as given in the table below.

Table VI The codeword Block 1 Block 2 0 x x 10 110 x 111 x

X indicates that the most likely mode should be used. For blocks for which the most probable mode m has not been used, a code number from 0 to 7 is transmitted indicating which of the remaining modes should be used. Since the remaining modes are ordered according to their numbers, the mode with the lower number precedes the mode with the higher number, when the code number q is received, the used prediction mode is:

q if q <m;

q + 1, otherwise.

The code number q is sent as a 3-bit code word, which is a binary representation of the corresponding code number.

The storage for the method in accordance with the present invention is based on the fact that the decoder for each combination of modes of blocks U and L must remember at most the ordered sets S and T, while S is the same for the whole combination of modes for blocks U and L .

As a result, the present invention provides methods, devices, and systems for encoding and decoding from a bit stream image information containing a plurality of image blocks using a plurality of spatial prediction modes for an intra block prediction mode. The spatial prediction mode of the current block (C) can be determined based on the set of obtained prediction modes obtained from the spatial prediction modes of a plurality of neighboring blocks (U, L) of the current block (C). Many of the n obtained prediction modes are grouped into a first group of m first prediction modes with m <n and a second group of second prediction modes. The first prediction modes have a higher probability of occurrence than the second prediction modes. While the first prediction modes are ordered according to the probability of their occurrence, the second prediction modes are arranged in a predetermined manner known to the decoder. Therefore, the ordered set is supplied to the decoding device in such a way as to enable the decoding device to determine the sequence of the second decoding modes based on the supplied ordered set. An ordered set is associated with one or more groups of spatial prediction modes of neighboring blocks (U, L) of the current block. In addition, the bitstream may include encoded information indicating which of the first prediction modes has the highest probability of occurrence when the obtained prediction mode with the highest probability of occurrence is selected for use in encoding the current block (C). The present invention also provides a computer program having pseudo codes intended for use in grouping a plurality of obtained prediction modes into first and second groups and in ordering the first prediction modes in accordance with the probability of occurrence among the first prediction modes.

Although the invention has been described with respect to a preferred embodiment, it will be understood by those skilled in the art that the above and other changes, omissions and deviations of the invention in form and detail may be made without departing from the scope of this invention.

Claims (31)

1. A method of encoding image information containing a plurality of image blocks into a bit stream using a plurality of spatial prediction modes for predicting a block in an internal mode, the method being that the spatial prediction mode of a current block is determined based on a plurality of obtained combinations of prediction modes obtained on based on spatial prediction modes of a plurality of neighboring blocks of the current block, characterized in that the prediction modes for each combination of prediction modes are divided into the first group and the second group, the first group of forecasting modes having m forecasting modes and the second group of forecasting modes (nm) forecasting modes, with n being the total number of available prediction modes and m less than n , and, in addition, the first group of prediction modes has a higher probability of occurrence than the second group of prediction modes, and encoding into the bitstream information indicating whether the spatial prediction mode of the current block belongs to said first group or to said second group. 2. The method according to claim 1, characterized in that they further streamline the first group of prediction modes in accordance with the probability of occurrence among the first group of prediction modes. 3. The method according to claim 1, characterized in that it further comprises a second group of prediction modes in a predetermined manner known to the decoding device. 4. The method according to claim 1, characterized in that it further encode in the bitstream information indicating which of the first group of prediction modes has the highest probability of occurrence when the obtained combination of prediction modes with the highest probability of occurrence is selected for use in encoding the current block. 5. The method according to claim 1, characterized in that the ordered set of prediction modes is known to the decoding device, so that the decoding device determines the sequence of the second group of prediction modes based on the ordered set of prediction modes. 6. The method according to claim 5, characterized in that the ordered set of prediction modes associated with more than one group of spatial prediction modes of many neighboring blocks of the current block. 7. The method according to claim 1, characterized in that the additionally obtained combinations of prediction modes are represented using many different integers, and the second group of prediction modes is ordered in accordance with a sequence of integers known to the decoding device. 8. The method according to claim 1, characterized in that the obtained combinations of prediction modes are represented using a variety of different integers and wherein the integers are arranged in an increasing sequence of integers to provide an ordered set of prediction modes, while additionally removing integers corresponding to the first group of prediction modes from the ordered set of prediction modes to provide a modified ordered set of prediction modes in such a way as to enable the decoding device to determine the sequence of the second group of prediction modes based on the modified ordered set of prediction modes. 9. The method according to claim 1, characterized in that one of the obtained combinations of prediction modes is selected for use in encoding the current block, while additionally provide a decoding device with information indicating an i-th mode among the second group of prediction modes if the selected mode of the combination of prediction modes is in the second group. 10. The method according to claim 1, characterized in that it further display the obtained combinations of spatial prediction modes of neighboring blocks to provide an additional combination of prediction modes of neighboring blocks, when necessary, using determining an additional prediction mode of the current block based on an additional combination of prediction modes of neighboring blocks and using displaying an additional prediction mode of the current block to obtain a spatial prediction mode of the current block. 11. The method according to claim 10, characterized in that the display of the additional prediction mode of said block is performed using the mirror image function. 12. An encoding device for encoding an image containing a plurality of image blocks using a plurality of spatial prediction modes for predicting a block in an internal mode, the spatial prediction mode of a current block being determined based on a plurality of obtained combinations of prediction modes that are obtained based on spatial prediction modes many neighboring blocks of the current block, characterized in that it contains means for separating prediction modes of each combination of prediction modes into a first group and a second group, the first group having m prediction modes and the second group having (nm) prediction modes, wherein n is the total number of available prediction modes and m is less than n, and, in addition, the first group of prediction modes has a higher probability of occurrence than the second group of prediction modes, and the second group of prediction modes is ordered in a specified way, it is known th decoding apparatus. means for encoding into the bitstream information indicating whether the spatial prediction mode of the current block belongs to said first group or to said second group. 13. The encoding device according to item 12, characterized in that it further comprises means responsive to said first group, designed to order the first group of prediction modes in accordance with the probability of occurrence among the first group of prediction modes to provide an ordered set of prediction modes. 14. The encoding device according to item 12, characterized in that it further comprises means for providing a decoding device with information indicating prediction modes in an ordered set of prediction modes that has the highest probability of occurrence. 15. The encoding device according to item 12, characterized in that it further comprises means configured to provide an additional prediction mode of neighboring blocks, when necessary, based on spatial prediction modes of neighboring blocks so that an additional prediction mode of the current block can be determined based on an additional prediction mode of neighboring blocks, and means responsive to the additional prediction mode of the current block, designed to provide a spatial prediction mode of the current block based on the display of the additional prediction mode of the current block. 16. A method of decoding from a bit stream image information containing a plurality of image blocks using a plurality of spatial prediction modes to predict a block in an internal mode, the spatial prediction mode of a current block being determined based on a plurality of obtained combinations of prediction modes obtained on based on spatial prediction modes of a plurality of neighboring blocks of the current block, characterized in that the prediction modes of each combination of prediction modes are divided into a first group and a second group, the first group having m prediction modes and the second group having (nm) prediction modes, wherein n is the total number of prediction modes available and m is less than n, and furthermore , the first group of prediction modes has a higher probability of occurrence than the second group of prediction modes, and information indicating whether the spatial prediction mode of the current block belongs to omyanutoy first group or said second group are encoded into the bitstream, and decode information from the bitstream indicating whether the spatial prediction mode of the current block belongs to the first group or the second group, so as to enable the decoding device to determine the spatial prediction mode of the current block based on the decoded information. 17. The method according to clause 16, characterized in that in addition the first group of prediction modes is ordered in accordance with the probability of occurrence among the first group of prediction modes. 18. The method according to clause 16, characterized in that in addition the second group of prediction modes have a predetermined manner known to the decoding device. 19. The method according to clause 16, wherein the additionally decoded information also indicates which of the first group of prediction modes has the highest probability of occurrence when a combination of prediction modes with the highest probability of occurrence is selected for use in encoding and decoding of the current block. 20. The method according to clause 16, characterized in that an additionally ordered set of prediction modes is supplied to the decoding device so as to enable the decoding device to determine the sequence of the second group of prediction modes based on the ordered set of prediction modes. 21. A decoder configured to decode image information containing a plurality of image blocks from a bit stream using a plurality of spatial prediction modes to predict a block in an internal mode, the spatial prediction mode of a current block being determined based on a plurality of obtained combinations of prediction modes obtained on based on spatial prediction modes of a plurality of neighboring blocks of the current block, in which the prediction modes of each combination of prediction modes are divided into a first group and a second group, the first group having m prediction modes and the second group having (nm) prediction modes, wherein n is the total number of prediction modes available and m is less than n, and furthermore , the first group of prediction modes has a higher probability of occurrence than the second group of prediction modes, characterized in that it contains means for responding to decoded information for determining whether the spatial prediction mode of the current block belongs to said first group or said second group, and means for selecting a spatial prediction mode based on the above definition. 22. The decoder according to item 21, wherein the prediction modes in the first group of prediction modes are ordered in accordance with the probability of occurrence among the first group of prediction modes. 23. The decoder according to item 21, characterized in that the second group of prediction modes is located in a predetermined manner known to the decoder, said decoder further comprising means for storing information pointing to the specified image. 24. The decoder according to item 21, wherein said determining means is further configured to determine, based on the decoded information, which of the first group of prediction modes has the highest probability of occurrence when the obtained combination of prediction modes with the mode having the highest probability of occurrence is selected for use in decoding the current block. 25. An image encoding system for encoding and decoding an image containing a plurality of image blocks using a plurality of spatial prediction modes for predicting a block in an internal mode, the spatial prediction mode of a current block being determined based on a plurality of obtained combinations of prediction modes obtained from prediction modes of many neighboring blocks of the current block, characterized in that it contains means for separating prediction modes of each combination of prediction modes into a first group and a second group, the first group having m prediction modes and the second group having (nm) prediction modes, wherein n is the total number of available prediction modes and m is less than n, and, in addition, the first group of prediction modes has a higher probability of occurrence than the second group of prediction modes, means for encoding group information into the bitstream indicating whether the spatial prediction mode of the current block belongs to said first group or to said second group, and means for decoding from the bit stream the said group information in such a way as to enable the decoding device to determine the spatial prediction mode of the current block based on said decoded information. 26. The image encoding system according A.25, characterized in that it further comprises means responsive to said first group, designed to streamline the first group of prediction modes in accordance with the probability of occurrence among the first group of prediction modes. 27. The image encoding system according A.25, characterized in that it further comprises means responsive to said second group, designed to arrange the second group of prediction modes in a predetermined manner known to the decoding device. 28. A software product stored in a computer-readable storage medium for use in an image encoding system for encoding image information containing a plurality of image blocks in a bit stream using a plurality of spatial prediction modes for predicting a block in an internal mode, the spatial mode prediction block is determined based on the set of obtained combinations of prediction modes that are received based on prediction modes of a plurality of neighboring blocks of said block, characterized in that it comprises pseudocode intended for use in the separation of prediction modes in each combination of prediction modes into the first group and the second group, the first group having m prediction modes and the second group having (nm) prediction modes, with n being the total number of prediction modes available and m less than n, and, in addition, the first group of prediction modes has a higher probability of occurrence than the second group of prediction modes, pseudocode intended for use in ordering the first group of prediction modes in accordance with the probability of occurrence among the first prediction modes. 29. The software product according to p. 28, characterized in that it further comprises pseudocode intended for use in the location of the second group of prediction modes in a predetermined manner known to the decoding device. Priority on points: 06/21/2002 according to claims 1-4, 7, 9-12, 15-19, 21-29; 07/09/2002 according to claims 5, 6, 8, 13, 14 and 20.

Images