WO2007063612A1

WO2007063612A1 - Dynamic image encoding device and dynamic image decoding device

Info

Publication number: WO2007063612A1
Application number: PCT/JP2006/303904
Authority: WO
Inventors: Maki Takahashi; Tomoko Aono
Original assignee: Sharp Kabushiki Kaisha
Priority date: 2005-11-30
Filing date: 2006-03-01
Publication date: 2007-06-07
Also published as: JP4855417B2; JPWO2007063612A1; JP2011147193A

Abstract

A dynamic image encoding device includes: a buffer memory for storing encoding information required for encoding respective blocks of an image; and an encoding unit having a data sort unit and a variable-length encoding unit for encoding a plurality of encoding information stored in the buffer memory in a unit of M × N (M and N are arbitrary integers). Thus, it is possible to provide a dynamic image encoding unit capable of performing encoding using a spatial correlation in a wide range.

Description

Specification

Video encoding device, video decoding device

Technical field

The present invention relates to a moving image encoding device and a moving image decoding device for realizing a high-efficiency encoding technique for moving image data.

Background art

[0002] As a conventional technique for realizing high-efficiency encoding of moving image data, the H.264 / AVC moving image encoding technology described in Non-Patent Document 1 can be cited.

[0003] A video encoding apparatus to which the above H.264 / AVC video encoding technology is applied will be described with reference to the drawings.

FIG. 13 is a block diagram of a conventional video encoding device.

In FIG. 13, reference numeral 1 is a frame memory for storing encoded images, reference numeral 2 is a prediction method control unit for determining a prediction method and prediction parameters used for an image to be encoded, and reference numeral 3 is , A prediction image generation unit that generates a prediction image from the encoded image stored in the frame memory 1, reference numeral 4 is a difference image generation unit that generates a difference image between the image to be encoded and the prediction image, and reference numeral 5 , An orthogonal transform unit that orthogonally transforms the difference image, code 6 is a quantization unit that quantizes the output data of the orthogonal transform unit 5 and outputs predicted residual data, and code 7 is input by an operation reverse to that of the quantization unit 6 Inverse quantization unit that performs inverse quantization of data, code 8 is an inverse transform unit that performs inverse orthogonal transform of input data by the reverse operation of orthogonal transform unit 5, and code 9 is the output data and prediction image of inverse transform unit 8 Image composition part to be composed, symbol 10 is quantity The data scanning unit, code 11, which rearranges the two-dimensional data output from the digitizing unit 6 into a one-dimensional data string by zigzag scanning, has a variable prediction method, prediction parameters, and prediction residual data for each element This is a variable-length encoding unit that performs long encoding.

[0006] The operation of the moving picture encoding apparatus having the above configuration will be described with reference to the operation flowchart of FIG. 14 in addition to FIG. Note that the video encoding apparatus performs the following processing for each macroblock consisting of a 16 × 16 pixel rectangular area.

[0007] STEP 1) The prediction method control unit 2 includes a prediction method used for the encoding target macroblock and A prediction parameter is determined (S101).

[0008] Here, as the prediction method, one of intra prediction and inter prediction (forward prediction, backward prediction, bidirectional prediction) is selected.

[0009] In addition, as a prediction parameter, a unit (macroblock division pattern) for prediction, in the case of intra prediction, which of a plurality of intra prediction methods is applied, or in the case of inter prediction, a motion vector A combination of reference images used for prediction is determined. Note that the prediction method and prediction parameter determination method uses the prediction image generation unit 3 that generates a prediction image with a combination of all prediction methods and prediction parameters, and has the highest correlation with the image to be encoded. However, since the amount of processing is enormous, the prediction methods and prediction parameters that are candidates for selection are limited in advance, and the prediction method and prediction parameter that obtains the most highly correlated prediction image among the candidates are selected. It is possible to use this method.

STEP 2) The variable length coding unit 11 performs variable length coding on the prediction scheme and the prediction parameter determined by the prediction scheme control unit 2 for each element (S102).

STEP 3) The predicted image generation unit 3 generates a predicted image of the macroblock based on the prediction method and the prediction parameter determined by the prediction method control unit 2 (S103).

[0012] STEP 4) The difference image generation unit 4 generates a difference image between the macroblock image and the predicted image (S104).

[0013] STEP 5) The orthogonal transform unit 5 performs orthogonal transform on the difference image of the macroblock for each 4x4 pixel block (S105).

STEP 6) The quantization unit 6 performs quantization on the 4 × 4 block data output from the orthogonal transform unit 5 (S106).

STEP 7) The inverse quantization unit 7 performs inverse quantization on the prediction residual data output from the quantization unit 6 (S107).

[0016] STEP 8) The inverse transform unit 8 performs inverse orthogonal transform on the 4x4 block data output from the inverse quantization unit 7 (S108).

STEP 9) The data scanning unit 10 performs zigzag scanning on the prediction residual data output from the quantization unit 6 and rearranges it into a one-dimensional data sequence (S109). [0018] STEP 10) The variable length coding unit 11 performs variable length coding on each element of the one-dimensional prediction residual data (S 110).

[Step 11] Steps 5 to 10 are repeated for all the 4 × 4 pixel blocks constituting the macro block (S11 11).

[0020] STEP 12) The image synthesis unit 9 synthesizes the output data of the inverse transformation unit 8 and the predicted image generated by the predicted image generation unit 3, encodes the synthesized image, and sets the encoded image as a subsequent image. It is stored in the frame memory 1 for use in encoding (S112).

[Step 13] Steps 1 to 12 are repeated for all macroblocks constituting the image to be encoded (S I 13).

Non-patent 乂 ffl ^ l: ITU-1 'Recommendation H.2 4: Advanced Video and oaing ror generic audiovisual services (2003)

Disclosure of the invention

[0022] As described above, in the conventional video encoding apparatus, encoding information (prediction method, prediction parameter, prediction residual data, etc.) is processed with a macroblock consisting of a rectangular area of 16x16 pixels as a processing unit. Is encoded. The macroblock size is constant both when encoding an image with a low spatial resolution and when encoding an image with a high spatial resolution. However, in general, the higher the spatial resolution, the higher the spatial correlation of the image to be encoded. Therefore, the code information also tends to show a high correlation between adjacent macroblocks.

[0023] Therefore, in a conventional video encoding apparatus, when encoding a high-resolution image in which a high correlation appears in a wider range than the macroblock size, a code that makes use of a spatial correlation in a wide range is used. There arises a problem that the conversion cannot be performed.

[0024] The present invention has been made in view of the above-described problems, and an object of the present invention is to encode a high-resolution image in which high correlation appears in a wider range than the macroblock size. However, it is an object of the present invention to provide a moving picture coding apparatus capable of performing coding that uses spatial correlation in a wide range.

[0025] In order to solve the above-described problem, the moving picture encoding apparatus according to the present invention is a moving picture encoding apparatus that divides an image into a plurality of blocks and encodes the blocks. Each temporary storage means for storing necessary encoding information and a plurality of pieces of encoding information stored in the temporary storage means are encoded with an MXN block (M and N are arbitrary integers) as a unit. It is characterized in that it is provided with sign key means for performing the above.

[0026] Specifically, the encoding means includes a data string conversion means for scanning the MXN block encoding information stored in the temporary storage means by a predetermined scanning procedure and converting it into a one-dimensional data string. A data string generating means for generating a plurality of one-dimensional data strings from the converted one-dimensional data string by a predetermined rule, and a data rearranging means for rearranging the generated one-dimensional data strings by a predetermined sorting rule. And variable length encoding means for performing variable length encoding on the rearranged one-dimensional data string.

[0027] Further, the data string conversion means may perform conversion into a one-dimensional data string in a different scanning order for each type of the encoded information.

[0028] Further, the data rearranging means may rearrange the one-dimensional data string using BWT (Burrows-Wheeler Transform).

[0029] When the spatial resolution of the encoding target is less than a predetermined threshold, the data rearranging means rearranges the one-dimensional data sequence for each Ml X N1 block, and the spatial resolution of the encoding target is predetermined. If the threshold is greater than or equal to the threshold, when the one-dimensional data sequence is rearranged for each M2 X N2 block, the block size relationship may satisfy Ml X NK M2 X N2.

[0030] The encoding information includes data indicating a prediction method to be applied to a moving image to be encoded, a prediction parameter used together with the prediction method, and applying the prediction method to a moving image to be encoded. It is desirable to have at least one of the required prediction residual data.

[0031] In order to solve the above-described problem, the moving picture decoding apparatus of the present invention is an MXN block (M, M, M) that decodes moving picture data encoded by the moving picture encoding apparatus. It is characterized by comprising decoding means for decoding encoded information encoded with N being an arbitrary integer) as a unit.

[0032] The decoding means includes variable-length decoding means for variable-length decoding encoded coding information, and data for generating a plurality of one-dimensional data sequences from the variable-length decoded encoding information Obtained by scanning the MXN block encoding information and the data sorting means that sorts the generated one-dimensional data strings according to a predetermined sorting rule, and the sorted one-dimensional data strings You may have a data string conversion means to convert to a one-dimensional data string.

[0033] The data string converting means may convert the data into a one-dimensional data string in a different reverse scanning order for each type of encoded information.

[0034] The data rearranging means may rearrange one-dimensional data strings using inverse BWT (Burrows-Wheeler Transform).

[0035] When the spatial resolution of the encoding target is less than a predetermined threshold, the data rearranging means rearranges the one-dimensional data sequence for each Ml X N1 block, and the spatial resolution of the encoding target is predetermined. If the threshold is greater than or equal to the threshold, it is desirable that the relationship between the block sizes satisfy M1 X NK M2 X N2 when rearranging the one-dimensional data sequence for each M2 X N2 block.

[0036] The encoding information includes data indicating a prediction scheme to be applied to a moving image to be encoded, a prediction parameter used together with the prediction method, and the prediction scheme applied to a moving image to be encoded. It is desirable to have at least one of the required prediction residual data.

[0037] As described above, the moving image encoding apparatus according to the present invention is a moving image encoding apparatus that divides and encodes an image into a plurality of blocks, and encodes information necessary for encoding each block. A temporary storage means for storing the code, and a code key means for performing a code key for a plurality of pieces of encoded information stored in the temporary storage means with an MXN block (M and N are arbitrary integers) as a unit; This enables encoding that takes into account the high correlation between adjacent blocks due to differences in spatial resolution. As a result, it is possible to perform wide encoding from low-resolution images to high-resolution images. There is an effect that can be done.

[0038] Further, the moving picture decoding apparatus according to the present invention can decode and reproduce the highly efficient moving picture encoded data encoded by the moving picture encoding apparatus according to the present invention. Brief Description of Drawings

FIG. 1 shows an embodiment of the present invention, and is a block diagram showing a main part configuration of a video encoding device. FIG.

FIG. 2 is a flowchart showing a flow of encoding processing by the video encoding apparatus shown in FIG. 1.

FIG. 3 (a) is a diagram illustrating an example of a 4 × 4 macroblock.

[FIG. 3 (b)] FIG. 3 (b) is a diagram showing a running pattern of the macroblock shown in FIG. 3 (a).

[FIG. 3 (c)] FIG. 3 (c) is a diagram showing a strike pattern of the macroblock shown in FIG. 3 (a).

[FIG. 3 (d)] is a diagram showing a running pattern of the macroblock shown in FIG. 3 (a).

[FIG. 3 (e)] is a diagram showing a running pattern of the macroblock shown in FIG. 3 (a).

[FIG. 3 (f)] FIG. 3 (a) is a diagram showing a strike pattern of the macroblock shown in FIG. 3 (a).

[Fig. 3 (g)] Fig. 3 (g) is a diagram showing a strike pattern of the macroblock shown in Fig. 3 (a).

FIG. 4 (a) is a diagram showing a running pattern of an 8 × 8 macroblock.

FIG. 4 (b) is a diagram showing a running pattern of an 8 × 8 macroblock.

FIG. 4 (c) is a diagram showing a scan pattern of an 8 × 8 macroblock.

FIG. 4 (d) is a diagram showing a scan pattern of an 8 × 8 macroblock.

FIG. 4 (e) is a diagram showing a scan pattern of an 8 × 8 macroblock.

FIG. 4 (f) is a diagram showing a scan pattern of an 8 × 8 macroblock.

5] (a) to (e) are flowcharts showing the flow of the data sorting process by the data sorting unit provided in the moving picture coding apparatus shown in FIG.

FIG. 6 (a) is a diagram showing an example of a 4 × 4 fixed size macroblock.

FIG. 6 (b) is a diagram showing a scanning pattern of the macroblock shown in FIG. 6 (a).

FIG. 6 (c) is a diagram showing a scanning pattern of the macroblock shown in FIG. 6 (a).

[Fig. 6 (d)] Fig. 6 (d) is a diagram showing a running pattern of the macroblock shown in Fig. 6 (a).

[Fig. 6 (e)] Fig. 6 (a) is a diagram showing a strike pattern of the macroblock shown in Fig. 6 (a).

[FIG. 6 (f)] FIG. 6 (a) is a diagram showing a strike pattern of the macroblock shown in FIG. 6 (a).

[Fig. 6 (g)] Fig. 6 (g) is a diagram showing a strike pattern of the macroblock shown in Fig. 6 (a).

FIG. 7 (a) is a diagram illustrating an example of a variable-size macroblock.

[Fig. 7 (b)] Fig. 7 (b) is a diagram showing a running pattern of the macroblock shown in Fig. 7 (a).

[FIG. 7 (c)] FIG. 7 (c) is a diagram showing a running pattern of the macroblock shown in FIG. 7 (a). FIG. 7 (d) is a diagram showing a scanning pattern of the macroblock shown in FIG. 7 (a).

FIG. 7 (e) is a diagram showing a scanning pattern of the macroblock shown in FIG. 7 (a).

FIG. 7 (f) is a diagram showing a scanning pattern of the macroblock shown in FIG. 7 (a).

[Fig. 7 (g)] Fig. 7 (g) is a diagram showing a strike pattern of the macroblock shown in Fig. 7 (a).

FIG. 8 is a diagram showing an example of a running pattern of predicted residual data.

FIG. 9 is a diagram illustrating an example of a variable-length encoding method by a variable-length encoding unit provided in the video encoding apparatus shown in FIG. 1.

FIG. 10, showing another embodiment of the present invention, is a block diagram showing a main configuration of a video decoding device.

FIG. 11 is a flowchart showing a flow of decoding processing of the video decoding device shown in FIG.

[FIG. 12] (a) to (g) are diagrams showing the flow of data reverse sorting processing by the data reverse sorting unit provided in the moving picture decoding apparatus shown in FIG.

FIG. 13 is a block diagram showing a main configuration of a conventional moving picture encoding device.

FIG. 14 is a flowchart showing a flow of encoding processing of the video encoding apparatus shown in FIG.

BEST MODE FOR CARRYING OUT THE INVENTION

[Embodiment 1]

A video encoding device according to a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a moving picture coding apparatus according to this embodiment.

In FIG. 1, reference numeral 12 denotes a buffer memory for temporarily storing encoding information (prediction method, prediction parameters, prediction residual data) of a plurality of macroblocks, and reference numeral 13 is recorded in a nota memory. Data sorting unit that rearranges the encoded information (prediction method, prediction parameter, prediction residual data) in a predetermined method, and code 14 is a variable length coding unit that performs variable length coding on the output result of the data sorting unit 13 It is. The MXN block (M, N is determined as follows) for a plurality of pieces of encoded information (prediction method, prediction parameter, prediction residual data) stored in the buffer memory 12 by the data sorting unit 13 and the variable length encoding unit 14. A code key unit (code key means) 100 is configured to perform coding with an arbitrary integer) as a unit.

[0042] It should be noted that the block having the same function as that of the conventional video encoding device (Fig. 13). Are the same in name and code and will not be described.

[0043] An outline of the operation of the moving picture coding apparatus having the above-described configuration will be described with reference to the flowchart of FIG. 2 in addition to FIG.

STEP 1) The prediction method control unit 2 determines a prediction method and a prediction parameter to be used for the codeh target macroblock (S1).

[0045] The prediction method and the prediction parameter determination method are the same as those of the conventional video encoding apparatus, and the description thereof will be omitted. The determined prediction method and prediction parameters are recorded in the buffer memory 12.

STEP 2) The predicted image generation unit 3 generates a predicted image of the macroblock based on the prediction method and the prediction parameter determined by the prediction method control unit 2 (S2).

[0047] STEP 3) The difference image generation unit 4 generates a difference image between the macroblock image and the predicted image (S3).

[0048] STEP 4) The orthogonal transform unit 5 performs orthogonal transform on the difference image of the macroblock for each 4x4 pixel block (S4).

[0049] STEP 5) The quantization unit 6 quantizes the 4x4 block data output from the orthogonal transform unit 5. The output prediction residual data is recorded in the buffer memory 12 (S5).

[0050] STEP 6) The inverse quantization unit 7 performs inverse quantization on the prediction residual data output from the quantization unit 6 (S6).

[0051] STEP 7) The inverse transform unit 8 performs inverse orthogonal transform on the 4x4 block data output from the inverse quantization unit 7 (S7).

[0052] STEP 8) Repeat steps 4 to 7 for all 4 X 4 pixel blocks constituting the macro block (S8).

STEP 9) The image composition unit 9 synthesizes the output data of the inverse transform unit 8 and the prediction image generated by the prediction image generation unit 3, and uses the generated image for the subsequent sign. Accumulate in memory 1 (S9).

STEP 10) STEP:! To 9 are repeated for all macroblocks constituting the image to be encoded (S 10).

[0055] STEP 11) The data sorting unit 13 is composed of MXN macroblocks (M and N are predetermined integers). Each of the prediction scheme, prediction parameters, and prediction residual data recorded in the buffer memory 12 is rearranged in a predetermined procedure in units of macroblock groups (Sl1). Details of the data sorting unit 13 will be described later.

[0056] STEP 12) The variable length coding unit 14 performs variable length coding for each of the prediction scheme, prediction parameters, and prediction residual data output from the data sorting unit 13 in units of macroblocks (S12). The details of the variable length code key unit 14 will also be described later.

STEP 13) Steps 1 to 12 are repeated until encoding of all the macroblock groups constituting the image to be encoded is completed (S13).

With the above procedure, the moving image data is encoded in the moving image encoder of the present embodiment. In the above description, in STEPs 11 and 12, the prediction method, the prediction parameters, and the prediction residual data have been described as being rearranged in units of macroblock groups, and variable length coding is performed. A part of (for example, prediction residual data) is input from the buffer memory 12 to the variable-length code unit 14 for each macroblock without passing through the data sorting unit 13, and in the same manner as in the prior art, It may be a configuration that performs sign coding.

Next, details of the data sorting unit 13 will be described.

[0060] The data sorting unit 13 converts the input data of a predetermined macroblock group into a data string that can be efficiently variable-length encoded using BWT (Burrows-Wheeler Transform).

[0061] BWT is a conversion for a one-dimensional data sequence. For example, a data sequence in which the same pattern repeatedly appears as "0,1,0,1, 0,1, ..." In this case, the output data string is a conversion having the property that elements having the same value are likely to be continuous compared to the input data string. Therefore, it is possible to perform efficient variable-length code by performing code assignment (for example, run-length code) collectively for elements having a series of identical values.

[0062] When handling data consisting of one element for one macroblock as in the prediction method, the macroblock group is processed as 4 X 4 macroblocks (MB0 to MB15) as shown in Fig. 3 (a), for example. Then, using one of the scissors patterns shown in Fig. 3 (b) to Fig. 3 (g) If the input data is converted to a one-dimensional data string, the same pattern is likely to appear repeatedly due to the correlation between adjacent macroblocks. A good sign can be achieved.

[0063] Similarly, if a macroblock group is processed as an 8 X 8 macroblock, if any one of the scissor patterns shown in Figs. 4 (a) to 4 (f) is used, It can be expected that elements with the same value in the output data string will be continuous.

[0064] Note that the strength of the correlation of the code key information between P-connected macroblocks varies depending on the spatial resolution of the image to be encoded, and therefore, the macro processor which is the processing unit of the data sorting unit 13 is used. For the size and running pattern of the group of images, it is desirable to perform a simulation of multiple images in advance and use the combination with the best performance.For example, when encoding a low-resolution image, When a macroblock group that processes 4x4 macroblocks and a high-resolution image is encoded, a macroblock group that processes 8x8 macroblocks has a high correlation between adjacent macroblocks due to differences in spatial resolution. It is possible to perform efficient coding considering the above. In addition, since the correlation level of the encoded information varies depending on the image to be encoded, a plurality of combinations are prepared in advance, and the combination used for encoding can be adaptively selected according to the image. It does not matter. In addition, since the strength of the correlation varies depending on the elements of the encoded information (prediction method, prediction parameter, prediction residual data), the configuration uses the macroblock group size and the scanning pattern for each element of the encoded information. It doesn't matter.

FIG. 5 shows a specific operation example of the data sorting unit. FIG. 5 (a) shows an example of input data to the data sorting unit 13 and here, an example in which data having element values 0 to 2 exists in each of 4x4 macroblocks. Yes. For example, each element value represents a prediction method, and 0: forward prediction, 1: backward prediction, 2: intra prediction.

[0066] STEP 1) By applying the scribing pattern shown in Fig. 3 (b) to Fig. 5 (a), the one-dimensional data string SO shown in Fig. 5 (b) is obtained.

[0067] STEP2) The SO element is shifted to the left by one element, and the first element is circulated to the tail to generate the one-dimensional data string S1. Similar operations are repeated to generate S1 to S15 (FIG. 5 (c)).

[0068] STEP 3) S0 to S15 are sorted in ascending order under the following conditions (FIG. 5 (d)). [0069] Two one-dimensional data strings Sa: = EaO, Eal, Ea2, ..., Eal5, Sb: = EbO, Ebl, ..., Eb

15 against

Condition l: Sa <Sb if EaO <EbO

Condition 2: Eai = Ebi (any i satisfying 0≤i <i <j), and if Eaj and Ebj, Sa is Sb STEP4) The last element of sorted S0 to S15 is the output data string in order. The SO rank in the sort result is output as additional information necessary for decoding (Fig. 5 (e)).

[0070] In the above description, the case where the data sort unit 13 handles data consisting of one element in one macroblock has been described. However, data having a plurality of elements in one macroblock can be handled in the same manner. it can.

[0071] For example, the prediction method, which is a prediction parameter in intra prediction, has a sub-matrix obtained by dividing one macro block into 4 X 4 sub-macro blocks (B0 to B15) as in the example shown in Fig. 6 (a). As shown in Fig. 6 (b) to Fig. 6 (g), considering the correlation between adjacent sub-macroblocks, as well as the force S in which data exists for each macroblock and the running pattern in units of macroblocks. Can be used.

The same applies to the case where there is a parameter for each sub-macroblock of variable size, such as designation of a motion vector reference image that is a prediction parameter for inter-frame prediction. For example, if there are variable-sized sub-macroblocks B0 to B4 as shown in Fig. 7 (a), the scan pattern itself is defined in the same way as in the fixed-size sub-block example (Fig. 6) and overlapped. If a sub-macroblock is scanned, the second and subsequent scans can be ignored. For example, in the case of the scan pattern in Fig. 7 (b), B0, Bl, B2, B3, B4 are output, and in the case of Fig. 7 (c), B0, B4, Bl, B2, B3 are output. In the case of 7 (d), output of B0, Bl, B4, B3, and B2 is satisfactory.

[0073] Also, in the case of prediction residual data, the same frequency component between adjacent blocks is considered to have a high correlation, so that four adjacent blocks (B0, B1) as shown in the example of FIG. , B4, B5), as shown in the order of the numerical values in the figure, the 16x4 element prediction residual data operation sequence is scanned from the low-frequency component to the high-frequency component, and the same scan is performed for the remaining blocks. It can be applied to the network.

Next, details of the variable length coding unit 14 will be described. Variable length coding of this embodiment As described above, the unit 14 performs efficient variable-length encoding by run-length encoding, using the property that elements having the same value in the data string output from the data sorting unit 13 are likely to continue. For run-length encoding, variable length encoding such as Goram code, Huffman code, and arithmetic code shown in Table 1 below may be used for encoding each element.

[0075] [Table 1]

As a specific example of the operation of the variable length encoding unit 14, a case where the output result of FIG. 5 used in the description of the data sorting unit 13 is variable length encoded will be described. For the sake of explanation, the Goram code in Table 1 is used for the sign of each element. In addition, when encoding using the Goram code of Table 1, the probability of occurrence of each element value is determined in advance so that the appearance probability decreases as the element value with the highest element value increases. Encoding becomes possible.

[0076] When the data string "2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 2, 1, 1, 1, 1" is represented by the element value and the repetition of the run length, 〃2, 4, 0, 7, 2, 1, 1, 4 ". To this, the codewords in Table 1 are assigned. However, since run length 0 does not exist, run length 1 is used as the element value. When codewords are assigned,

"011, 00100, 1, 00111, 011, 1, 010, 00100 〃. Furthermore, by assigning 0, which is the order of SO sorting results, to Table 1,

A total of 27 bits “011, 00100, 1, 00111, 011, 1, 010, 00100, 1” is the output of the variable length encoding unit 14.

[0077] For reference, as in the conventional technique, the input data string S0 in FIG. 5 = S0 = 〃0, 0, 1, 2, 0, 0, 1, 2, 2, 0, 0, 1, 2, 0, 0, 2 When each variable of 2〃 is individually encoded with variable length using Table 1, " 1, 1, 010, 011, 1, 1, 010, 011, 1, 1, 010, 011, 1, 1, 011, 011〃 in total, 32 bits are required. It can be seen that the variable-length coding based on the 5 bit coding efficiency is better.

Further, by predicting the element values that appear, it is possible to expect further improvement in code efficiency.

[0079] For example, the encoding procedure using prediction is as follows.

[0080] STEP 1) Predicted value candidate strings are initialized in descending order of element values. The predicted value candidate string indicates that the closer to the head, the higher the appearance probability, and the closer to the tail, the lower the appearance probability.

[0081] STEP 2) A search is performed for the number of the leading element value of the input data string that appears in the predicted value candidate string.

[0082] STEP 3) Output the index and run length_1 of the predicted value obtained by the search, and variable-length encode each.

STEP 4) The prediction candidates from the beginning of the prediction value candidate sequence to the prediction value index obtained in STEP 2 are moved to the end of the prediction candidate sequence.

[0084] STEP 5) Repeat steps 2 to 4 until all data strings in the input data string are encoded.

[0085] FIG. 9 shows that the above-described data string “2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 2, 1, 1, 1, 1〃 This is an example of encoding, and 〃1, 00100, 010, 00111, 1, 1, 1, 00100 "are obtained as output results. If the encoding result of 0, which is the order of the SO sorting results, is calculated, it becomes 23 bits in total, "1, 00100, 010, 00111, 1, 1, 1, 00100, 1". It can be seen that efficient coding is performed compared to run-length coding.

[0086] As described above, in the moving picture coding apparatus according to the present embodiment, a conventional moving image is obtained by performing variable-length coding using a correlation between sub-macroblocks and a macroblock. Encoding efficiency can be increased compared to the encoding device.

[Embodiment 2]

Next, as another embodiment of the present invention, a moving picture decoding apparatus for decoding encoded data encoded by the moving picture encoding apparatus of the first embodiment will be described.

FIG. 10 is a block diagram showing a video decoding device according to the present embodiment.

In FIG. 10, reference numeral 15 denotes a variable length code decoding unit that performs variable length decoding of encoded data; Reference numeral 16 denotes a data reverse sorting unit that rearranges the one-dimensional data sequence obtained by variable length decoding by a predetermined operation and restores the data to data for each macroblock.

Note that the blocks having the same functions as those of the moving picture coding apparatus according to Embodiment 1 have the same names and symbols, and the description thereof is omitted.

Next, the operation of the video decoding apparatus according to the present embodiment will be described below with reference to the operation flow of FIG. 11 in addition to FIG.

[0091] STEP 1) The variable-length code decoding unit 15 performs variable-length decoding on the encoded data, and in the video image encoding device shown in Fig. 1, the output state prediction method, prediction parameters, and prediction residual data of the data sorting unit 13 To restore. Details of the variable-length code decoding unit 15 will be described later (S1).

STEP 2) The data reverse sort unit 16 rearranges the one-dimensional data sequence (encoded information) input from the variable length decoding unit 15 into a two-dimensional data sequence by a predetermined procedure and records it in the buffer memory 12. (S2). The data storage format recorded in the buffer memory is the same as the data storage format in the buffer memory in the moving picture encoding apparatus of Embodiment 1, and the details of the operation of the data reverse sort unit 16 will be described later.

[0093] STEP 3) Steps 1 and 2 are repeated until the data of all the macroblock groups constituting the screen are restored (S3).

[0094] STEP 4) The inverse quantization unit 7 receives the prediction residual data from the buffer memory 12, and performs inverse quantization (S4).

[0095] STEP 5) The inverse transform unit 8 performs inverse orthogonal transform on the 4x4 block data output from the inverse quantization unit 7 (S5).

[0096] STEP 6) Steps 4 to 5 are repeated for all the 4x4 pixel blocks constituting the macro block (S6).

STEP7) The predicted image generation unit 3 reads the prediction method and prediction parameters from the nother memory 12, and generates a predicted image based on the decoded images stored in the frame memory 1 (S7).

STEP 8) The image synthesis unit 9 synthesizes the output data of the inverse transformation unit 8 and the prediction image generated by the prediction image generation unit 3, and uses the generated image for subsequent decoding in the frame memory 1 And output to the outside (S8).

STEP 9) Repeat steps 4 to 8 for all macroblocks constituting the image (S9).

Next, the variable length code decoding unit 15 will be described in detail.

The operation of the variable length code decoding unit 15 performs the reverse operation of the variable length coding unit 14 in the first embodiment.

[0099] When decoding the encoded data that has been encoded by the simple run length described above, the input encoded data string "011, 00100, 1, 00111, 011, 1, 010, 00100, 1" is described above.表 2, 3, 0, 6, 2, 0, 1, 3, 0〃 are obtained, and the restoration of the run and the order of SO are separated, and the data sequence "2, 2 , 2, 2, 0, 0, 0, 0, 0, 0, 0, 2, 1, 1, 1, 1 ", SO rank: 0 completes decoding.

[0100] In addition, when encoding using a prediction value is performed, decoding is performed according to the following procedure.

[0101] STEP1) The candidate sequence of predicted values is initialized in descending order of element values, similar to encoding.

[0102] STEP2) Decode the predicted index and run length based on the code assignment in Table 1.

[0103] STEP 3) The value indicated by the prediction value index obtained in STEP 2 in the prediction candidate string is decoded as an element value.

[0104] STEP4) Update prediction candidate sequence based on prediction value index as in encoding

[0105] STEP 5) Repeat steps 1 to 4 until all elements are decoded.

[0106] STEP 6) The order of SO is decoded based on the code assignment shown in Table 1 to complete the decoding.

Next, the operation of the data reverse sort unit 16 will be described with reference to FIG. 12 in the case of restoring the example (FIG. 5) used in the operation description of the data sort unit described above. In the following description, it is assumed that the data reverse sort unit 16 is input with the data string shown in FIG. This input data is the same as the output result shown in FIG.

STEP1) The last element value of S0 to S15 in Fig. 5 (d) is restored by assigning the input data sequence in the order of rank 0 to 15 (Fig. 12 (b)).

[0108] In addition, the top elements of S0 to S15 in Fig. 5 (d) can sort the input data string in ascending order. And obtained. It should be noted that the order of S1 to S15 other than SO is unknown at this point.

[0109] STEP2) The last elements of S0 to S15 in Fig. 12 (c) are assumed to be e0, el, ·, and el5 from the upper element (in order from the top of the figure). Search for the first element of S0 to S15 having the same element value in order from e0 to el5, and assign the same symbol. When there are a plurality of identical element values, the same symbol is assigned to the upper element.

[0110] STEP3) Shift all elements other than the last of S0 to S15 back by one, and move the last element to the beginning (Figure 12 (d)). This operation reveals which element follows each element of e0, el, ···, el5.

[0111] STEP4) The third to fourteenth elements of SO are determined based on the information obtained in Fig. 12 (d) (Fig. 12 (e)).

[0112] STEP5) Replace the symbol assigned to each element of SO with the corresponding element value (Fig. 12 (f)).

[0113] STEP 6) Using the scanning pattern used in the moving picture encoding apparatus of Embodiment 1, one-dimensional data is converted back to two-dimensional data, and each element value of the encoding information in the MxN macroblock is restored. And output to the buffer memory 12 (FIG. 12 (g)).

[0114] The encoded data generated by the video encoding apparatus of the first embodiment can be decoded by the operation described above.

[0115] In the above description, as the macro block size and the sub macro block size,

Although the same size as the conventional technology was used, a configuration using a different size is acceptable. Also, the macroblock group size may be different for each element of the encoded information (prediction method, prediction parameter, prediction residual data).

[0116] Also, the moving image encoding method of the present invention is a moving image encoding method for dividing and encoding an image into a plurality of blocks, and temporarily encoding information necessary for encoding each block. And a plurality of pieces of encoded information stored in the previous step, the code is encoded using 1 ^ 1 ^ blocks (1 ^, N are arbitrary integers) as a unit. It is characterized by including a tape.

[0117] The encoding step includes a data string conversion step of scanning the temporarily stored encoding information of the MXN block according to a predetermined scanning procedure and converting it into a one-dimensional data string. And a data column generation step for generating a plurality of one-dimensional data strings from the converted one-dimensional data string by a predetermined rule, and a data arrangement for rearranging the generated one-dimensional data strings by a predetermined sorting rule. And a variable length code step for performing variable length encoding on the rearranged one-dimensional data string.

[0118] The moving image decoding method of the present invention is the moving image decoding method for decoding moving image data encoded by the moving image encoding method, wherein an MXN block (M and N are arbitrary integers) is a unit. It is characterized by including a step of decoding the encoded information.

[0119] The moving image decoding method of the present invention is the moving image decoding method for decoding moving image data encoded by the moving image encoding method, wherein an MXN block (M and N are arbitrary integers) is a unit. A decoding step for decoding the encoded information, wherein the decoding step includes a variable-length decoding step for variable-length decoding the encoded information, and a plurality of encoding information obtained from the variable-length decoding. A data string generation step that generates a one-dimensional data string, data that sorts multiple generated one-dimensional data strings according to a predetermined sorting rule, a sorting step, and the sorted one-dimensional data string are converted to an MXN block. And a data string conversion step for converting the encoded information into a one-dimensional data string obtained by scanning.

[0120] The present invention is not limited to the above-described embodiments, but can be variously modified within the scope of the claims, and can be obtained by appropriately combining technical means disclosed in different embodiments. Such embodiments are also included in the technical scope of the present invention.

[0121] Finally, each block of the moving image encoding apparatus, in particular, the data sort unit 13 and the variable length encoding unit 14 may be configured by hardware logic, or software using a CPU as follows. Can be realized by.

[0122] That is, the moving image encoding apparatus includes a CPU (central processing unit) that executes instructions of a control program that realizes each function, a ROM (read only memory) that stores the program, and a RAM that expands the program. (random access memory), and a storage device (recording medium) such as a memory for storing the program and various data. The object of the present invention is to provide a moving image encoding device that is software that realizes the above-described functions. A recording medium in which the program code (execution format program, intermediate code program, source program) of the control program is recorded so as to be readable by a computer is supplied to the above moving image encoding device, and the computer (or CPU or MPU) records it. It can also be achieved by reading and executing the program code recorded on the medium.

[0123] The recording medium includes, for example, a tape system such as a magnetic tape and a cassette tape, a magnetic disk such as a floppy (registered trademark) disk Z hard disk, and an optical disk such as CD-ROM / MOZ MDZDVD / CD-R. A disk system, a card system such as an IC card (including a memory card) / optical card, or a semiconductor memory system such as a mask ROMZEPROM / EEPROMZ flash ROM can be used.

[0124] Further, the moving picture coding apparatus may be configured to be connectable to a communication network, and the program code may be supplied via the communication network. The communication network is not particularly limited. For example, the Internet, intranet, extranet, LAN, ISDN, VAN, CATV communication network, virtual private network, telephone line network, mobile communication network, satellite communication A net or the like is available. Also, the transmission medium constituting the communication network is not particularly limited. For example, IEEE1394, USB, power line carrier, cable TV line, telephone line, ADSL line, etc. ooth (registered trademark), 802.11 wireless, HDR, mobile phone network, satellite line, terrestrial digital network, etc. can also be used. The present invention can also be realized in the form of a computer data signal embedded in a carrier wave, in which the program code is embodied by electronic transmission.

Industrial applicability

[0125] The present invention can also be applied to a device that requires high-efficiency encoding of high-definition moving image data, such as a recording / playback device such as an HDTV.

Claims

The scope of the claims

[1] In a video coding apparatus that divides and codes an image into a plurality of blocks,

Temporary storage means for storing encoding information necessary for encoding each block, and a plurality of encoding information stored in the temporary storage means, MXN blocks (M and N are arbitrary integers) A moving picture coding apparatus comprising coding means for performing coding as a unit.

[2] The encoding means includes:

Data string conversion means for scanning the code information of the M X N block stored in the temporary storage means and converting it into a one-dimensional data string by a predetermined scanning procedure;

A data string generation means for generating a plurality of one-dimensional data strings from the converted one-dimensional data string according to a predetermined rule;

A data rearranging means for rearranging a plurality of generated one-dimensional data strings according to a predetermined rearrangement rule;

2. The moving picture coding apparatus according to claim 1, further comprising: variable length coding means for performing variable length coding on the rearranged one-dimensional data string.

[3] The data string conversion means includes:

3. The moving picture coding apparatus according to claim 2, wherein conversion into a one-dimensional data string is performed in a different running order for each type of coding information.

[4] The data sorting means includes:

3. The moving picture coding apparatus according to claim 2, wherein the one-dimensional data sequence is rearranged using BWT (Burrows-Wheeler Transform).

[5] The data sorting means includes:

When the spatial resolution of the encoding target is less than a predetermined threshold, the one-dimensional data sequence is rearranged for each Ml X N1 block,

When the spatial resolution of the encoding target is greater than or equal to a predetermined threshold, when rearranging the one-dimensional data sequence for each M2 X N2 block,

The block size relationship satisfies Ml X Nl <M2 X N2. The moving image encoding apparatus described.

[6] The encoding information includes data indicating a prediction scheme to be applied to a moving image to be encoded, a prediction parameter to be used together with the prediction method, and applying the prediction method to a moving image to be encoded. 6. The moving picture coding apparatus according to any one of claims 1 to 5, wherein at least one piece of predicted residual data is obtained.

[7] In the moving picture decoding apparatus for decoding moving picture data encoded by the moving picture encoding apparatus according to claim 1,

A moving picture decoding apparatus comprising decoding means for decoding encoded information encoded with M X N blocks (M and N are arbitrary integers) as a unit.

[8] In the moving image decoding apparatus for decoding moving image data encoded by the moving image encoding apparatus according to any one of claims 2 to 6,

Decoding means for decoding encoded information encoded with M X N blocks (M and N are arbitrary integers) as a unit;

The decoding means includes

Variable-length decoding means for variable-length decoding encoded encoding information;

Data sequence generation means for generating a plurality of one-dimensional data sequences from variable length decoded encoded information;

A moving picture decoding apparatus comprising: data string conversion means for converting the rearranged one-dimensional data string into a one-dimensional data string obtained by scanning the code information of the M X N block.

9. The moving picture decoding apparatus according to claim 8, wherein the data string converting means converts the data into a one-dimensional data string in a different reverse scanning order for each type of the encoded information.

10. The moving picture decoding apparatus according to claim 8, wherein the data rearranging means rearranges the one-dimensional data sequence using inverse BWT (Burrows-Wheeler Transform).

[11] The data rearranging means includes:

When the spatial resolution to be encoded is less than a predetermined threshold, the Ml X N1 block Sort the one-dimensional data string,

9. The moving picture decoding apparatus according to claim 8, wherein the block size relationship satisfies Ml X Nl <M2 X N2.

The encoding information includes data indicating a prediction method to be applied to a moving image to be encoded, a prediction parameter used together with the prediction method, and a prediction parameter for the moving image to be encoded. 12. The moving picture decoding apparatus according to any one of claims 7 to 11, wherein at least one of the prediction residual data obtained by applying the method is at least one.