JPH11275586A - Coder, coding method and record medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To extend a motion compensation range in the case that an image is divided and parallel processing is applied to divided images. SOLUTION: Image data in a area A are inputted to an upper area F prediction memory 51 and a half-pell processing section 60 applies half-pell processing to the data and the result is fed to a selector 57. Image data in an area B are inputted to an F prediction memory 52 and a half-pell processing section 61 applies half-pell processing to the data and the result is fed to the selector 57. Image data in an area C are inputted to a lower area F prediction memory 53 and a half-pell processing section 62 applies half-pell processing to the data and the result is fed to the selector 57. The selector 57 selectively outputs image data in a prescribed area as a predicted image signal based on an area select signal fed from a select signal generating section 59.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding apparatus, and more particularly, to an encoding apparatus that divides an image into a plurality of regions and performs encoding on each region in parallel.

[0002]

2. Description of the Related Art Conventionally, as a method of compressing and encoding moving image data, MPEG (Moving Picture Experts Gr.
oup). This encoding method differs in processing capability required for encoding processing according to the spatial resolution of an image. Therefore, for example, when encoding an image signal of a high-definition moving image such as a high-definition video, NTSC
An extremely high processing capability is required as compared with the case of encoding the image signal of the system.

Therefore, it has been considered that an image is divided into a plurality of areas (areas), and encoding processing is performed for each area in parallel.

[0004]

However, in image processing that requires high-speed processing, when an image is divided into a plurality of areas and parallel processing is performed, the motion compensation range is limited to each divided area. There was a problem to be done.

The present invention has been made in view of such a situation, and expands a motion compensation range to a desired range by making it possible to refer to image data of an adjacent area from a predetermined area. Is to be able to do.

[0006]

According to a first aspect of the present invention, there is provided an encoding apparatus which divides an image into a plurality of regions and encodes image data for each region. Motion vector detection means for detecting a motion vector, first storage means for storing image data of the area for each area, and second storage means for storing image data of another area adjacent to the area And image data of a certain area stored in the first storage means and image data of another area stored in the second storage means, based on the motion vector detected by the motion vector detection means. Selecting means for selectively outputting image data of a predetermined area of the selected area, using the image data output from the selecting means,
The motion compensation is performed. The area corresponding to the image data stored in the second storage means is the first area.
May be located at least one of up, down, left, and right with respect to the area corresponding to the image data stored in the storage means. Further, the second storage means includes a memory for storing image data of an area adjacent to an area above the area corresponding to the image data stored in the first storage means, and an image data stored in the first storage means. And a memory for storing image data of an area adjacent to the lower side of the area corresponding to. The second storage means includes a memory for storing image data of an area adjacent to the left side of the area corresponding to the image data stored in the first storage means, and an image data stored in the first storage means. And a memory for storing the image data of the area adjacent to the right side of the area corresponding to. Further, the first storage means and the second storage means are constituted by a memory for performing forward inter-frame predictive coding and a memory for performing reverse inter-frame predictive coding. be able to. Further, a half-pel unit that performs a half-pel process on the image data stored in the first storage unit and the second storage unit is further provided, and the half-pel unit is provided for each image data of each area. You can do so. Further, a half-pel unit that performs a half-pel process on the image data output from the selection unit may be further provided. An encoding method according to claim 8, wherein the image is divided into a plurality of regions, and the image data is encoded for each region. The image data of the area is stored, and the image data of another area adjacent to the area is stored. Based on the detected motion vector, the stored image data of the area and the image of the other area are stored. Image data of a predetermined area of the data is selectively output, and motion compensation is performed using the image data. A recording medium according to a ninth aspect is characterized by recording a program capable of executing the encoding method according to the eighth aspect. In the encoding device, the encoding method, and the recording medium according to the present invention, an image is divided into a plurality of regions, a motion vector of image data is detected, and for each region, image data of the region is stored. The image data of another area adjacent to the area is stored, and based on the detected motion vector, the stored image data of the area and the image data of a predetermined area among the image data of the other area are stored. The image data is selectively output, and motion compensation is performed using the image data.

[0007]

FIG. 1 is a block diagram showing a configuration example of an embodiment of an encoding apparatus according to the present invention. The embodiment shown in FIG. 1 includes three encoding units 1 to 3 for simplicity of description. Hereinafter, only the encoding unit 1 will be described, but the encoding units 2 and 3 have the same configuration as that of the encoding unit 1. Therefore, the description thereof will be appropriately omitted. The encoder 1 performs an encoding process on the area A. The encoder 2 performs an encoding process on the area B, and the encoder 3
Is configured to perform an encoding process on the area C. Here, the image is divided into three in the vertical direction, of which the uppermost area is the area A, the middle area is the area B, and the lowermost area is the area C.

[0008] ME (Motion Estima) constituting the encoding unit 1
Option 11 receives the video input signal 101, calculates a motion vector, and outputs a corresponding motion vector signal. DCT (discrete cosine transfo
rm) 13 performs a discrete cosine transform process on the signal supplied from the subtracter 19 and outputs the signal. Quantizer 1
4 is connected to the DCT 13 and the encoding control unit 1A,
The signal subjected to the discrete cosine transform supplied from the CT 13 is quantized by the number of steps corresponding to the code amount control signal output from the encoding control unit 1A, and is output.

The inverse quantizer 15 is connected to the quantizer 14 and the encoding control unit 1A, and converts a signal output from the quantizer 14 according to a code amount control signal output from the encoding control unit 1A. It is inversely quantized by the number of steps and output. IDCT (inversediscrete cosine transform)
Reference numeral 16 is connected to the inverse quantizer 15, and performs an inverse discrete cosine transform on the signal inversely quantized by the inverse quantizer 15, and outputs the image signal as an image signal. The adder 17 has an IDCT
16 is connected to the selector 18 and the selector 18, and generates the decoded image signal 105 by adding the predicted image signal output from the selector 18 and the inverse discrete cosine transformed image signal output from the IDCT 16. It has been made to be.

The prediction memory unit 12 includes an encoding control unit 1A,
The decoder 17 is connected to the adder 17 and the ME 11 and stores the decoded image signal 105 generated under the control of the encoding control unit 1A and supplied from the adder 17. Then, a predicted image signal stored at an address corresponding to the motion vector signal supplied from the ME 11 is output.

The image selector 18 is connected to the prediction memory unit 12 and the encoding control unit 1A, and calculates two prediction image signals output from the prediction memory unit 12 and a prediction image obtained by calculating an average of them. Of the signals, the video input signal 101
And a predicted image signal that minimizes the MAE (sum of absolute differences) between the two is selected and output.

The subtractor 19 is connected to the image selector 18 and the DCT 1
3, and calculates the difference between the video input signal and the predicted image signal supplied from the image selector 18 and supplies the difference to the DCT 13.

Next, a detailed configuration example of the prediction memory units 12, 22, 32 will be described. These prediction memory units 1
2, 22, and 32 can be realized by, for example, a combination of a memory, a selector, and a select signal generation unit for performing a comparison process.

FIG. 2 shows the prediction memory unit 12 shown in FIG.
It is a block diagram which shows the detailed example of a structure of 22 and 32. Since the prediction memory units 12, 22, and 32 have the same configuration, the prediction memory unit 22 will be described here as an example. Further, the F (Forward) prediction memories 51, 52, 53 and the selector 57 of FIG. 2 and the B (Backward) prediction memories 54, 55, 56,
Also, the selector 58 has the same configuration, so that only the F block will be described here, and the description of the B block will be appropriately omitted.

As shown in FIG. 2, the prediction memory unit 22
Are upper area F prediction memory 51, F prediction memory 52, lower area F prediction memory 53, half-pel processing units 60 and 6,
1, 62, selector 57, upper area B prediction memory 54,
It comprises a B prediction memory 55, a lower area B prediction memory 56, half-cell processing units 63, 64, 65, a selector 58, and a select signal generation unit 59.

In the upper area F prediction memory 51, the decoded image signal 501 supplied from the upper adjacent area A is written. Then, output from the select signal generation unit 59,
The motion vector signal 202 supplied from the ME 21 in FIG.
And a predicted image signal 507 corresponding to the read address signal 517 is output via the half-pel processing unit 60.

Similarly, the decoded image signal 502 is input to the F prediction memory 52 and written. Then, a read address signal 517 corresponding to the motion vector signal 507
Of the predicted image signal 508 corresponding to
Output via. Further, the decoded image signal 503 is supplied to the lower area F prediction memory 53 from the lower adjacent area C and written therein. Then, a predicted image signal 509 corresponding to the read address signal 517 corresponding to the motion vector signal 507 is output via the half-pel processing unit 62.

The selector 57 includes F prediction memories 51 and 5
2, 53, and selects one of the predicted image signals 507, 508, 509 from the corresponding half-pel processing units 60, 61, 62 in accordance with the area select signal 513. , Which are output as predicted image signals 515.

Next, the operation of the embodiment shown in FIG. 1 will be described with reference to FIGS. 1, 2 and 3.

When the video input signal 101 is supplied to the subtractor 19, the subtracter 19 calculates the difference between the video input signal 101 and the predicted image signal supplied from the selector 19, and outputs the image signal as the operation result to the DCT 13 To supply. DCT
13 is for the image signal supplied from the subtractor 19,
A discrete cosine transform process is performed, and the result is supplied to the quantizer 14. The quantizer 14 quantizes the discrete cosine transformed signal supplied from the DCT 13 by the number of steps according to the code amount control signal supplied from the encoding control unit 1A, and outputs the quantized signal.

The inverse quantizer 15 inversely quantizes the signal supplied from the quantizer 14 by the number of steps according to the code amount control signal supplied from the encoding control unit 1A, and outputs the result.
The IDCT 16 performs an inverse discrete cosine transform process on the signal subjected to the inverse quantization process by the inverse quantizer 15, and then supplies the signal to the adder 17. The adder 17 includes a selector 18
The decoded image signal 105 is generated by adding the predictive image signal supplied from the IDCT 16 and the image signal subjected to the inverse discrete cosine transform process supplied from the IDCT 16, and outputs the decoded image signal 105.

The decoded image signal 105 generated and output in the adder 17 is supplied to the prediction memory unit 12 and also to the prediction memory unit 22 in the next lower B area at the same timing. However, since the A area corresponding to the encoding unit 1 is the uppermost area in this example,
There is no supply to the adjacent upper area.

Similarly, the decoded image signal 205 output from the adder 27 of the encoding unit 2 is supplied to the prediction memory unit 22, and the prediction memory unit 12 in the A area on the adjacent upper
Also, it is supplied at the same timing to the prediction memory unit 32 in the area C in the next lower area.

The decoded image signal 305 output from the adder 37 of the encoding unit 3 is supplied to the prediction memory unit 32 and also to the prediction memory unit 22 in the next upper B area at the same timing. Is done. However, in this case, the C area corresponding to the encoding unit 3 is the lowest area, so that the supply to the adjacent lower area is not performed.

Depending on the relationship between the division method and the motion compensation range, there may be access from area A to area C and vice versa.

The prediction memory unit 12 includes an encoding control unit 1A
Of the motion vector signal 105 calculated based on the video input signal 101 by the ME 11 is written.
Then, a predicted image signal of an address corresponding to 2 is output. The image selecting unit 18 selects and outputs a predicted image signal having a minimum MAE (sum of absolute differences) between the signal supplied from the prediction memory unit 12 and the video input signal 101.

Next, the operation of the prediction memory unit 22 shown in FIG. 2 will be described with reference to the timing chart of FIG.

Here, a decoded image signal 501 input from the upper adjacent area A, a decoded image signal 502 in the B area,
The decoded image signal 503 input from the lower neighboring area C is
It is assumed that they are input at the same timing. Then, at the same timing, the F prediction memory 5
Assume that data is written to 1, 52, and 53.

As shown in FIG. 2, the motion vector signal 2
02 corresponding to the read address signal 517 corresponding to
The three F prediction memories 51, 52, and 53 are read out at the same timing, and the corresponding half-pel processing units 60, 6
The signals are supplied to the selector 57 via the respective channels 1 and 62.

The selector 57 receives the area select signal 51
In accordance with 3, one of the predicted image signals 507, 508, 509 supplied via the half-pel processing units 60, 61, 62 is selected and output as a predicted image signal 515.

The decoded image signal 504 input from the upper adjacent area A, the decoded image signal 505 of the B area, and the decoded image signal 505 input from the lower adjacent area C are input at the same timing. Is assumed. Then, at the same timing, the B prediction memory 54,
Assume that it is written to 55,56.

Similarly, in response to the read address signal 517 corresponding to the motion vector signal 202, the three B prediction memories 54, 55 and 56 are read out at the same timing, and the corresponding half-pel processing units 63 , 64 and 65, respectively, to the selector 58.

The selector 58 outputs the area select signal 51
According to 4, one of the predicted image signals 510, 511, 512 supplied via the half-pel processing units 63, 64, 65 is selected and output as the predicted image signal 516.

Specifically, for example, as shown in FIG. 3, in this example, since the value of the motion vector signal is 0 at the addresses 0 to 127, the B prediction memory 55 corresponding to the B area is The selected image signals are output from the selector 58 as predicted image signals B0 to B127. On the other hand, at addresses 128 to 255, since the value of the motion vector signal is -127 in the vertical direction, the prediction memory 1 in the A area
2 is used as reference data. Therefore, the upper area B prediction memory 54, that is, the A area is selected, and the predicted image signals A128 to A2 are output from the selector 58.
55 is output.

That is, the image data of each of the A area, the B area, and the C area corresponds to the addresses 0 to 255,
Above the area corresponding to the addresses 0 to 127 in the B area, the area corresponding to the addresses 127 to 255 in the A area is adjacent.

The following effects can be obtained in the coding apparatus having the above-described configuration.

A first effect is that the motion compensation range can be expanded. That is, when a motion compensation range exceeding one divided area is required, a memory for recording a newly added area as a motion compensation range for that area is used as a prediction (reference) memory in each area. Just add it. That is, a memory for storing image data of another area required for reference by one divided area is added. As a result, it is possible to eliminate the limitation of the motion compensation range due to the division processing, and to expand the motion compensation range. The reason is that the area is divided for each memory, and the area is selected by the motion vector.

The second effect is that, for each divided area,
There is no need to store the frame image data of the entire one screen in a memory or the like. That is, the memory can be reduced. The reason is that the encoding apparatus of the present invention is configured to store only image data of an area necessary for prediction.

As a third effect, the speed of image processing can be increased. The reason is that parallel processing can be performed without increasing the operation speed.

As a fourth effect, it is possible to enlarge the image processing screen. The reason is that parallel processing can be performed without increasing the operation speed.

A fifth effect is that the coding apparatus of the present invention can be applied to a divided high-efficiency coding apparatus and further to a divided high-efficiency decoding apparatus. The reason is that the encoding device of the present invention simply divides an image into slices and performs parallel processing, so that the motion compensation circuit is used in both the encoding device and the decoding device in the same manner. Because it can be.

FIG. 4 is a block diagram showing a configuration example of another embodiment of the prediction memory unit of the embodiment of the encoding device shown in FIG. In the prediction memory unit shown in FIG. 4, the half-pel processing unit 60 in the prediction memory unit 22 shown in FIG. 2 is arranged after the selector 57, and performs half-pel processing on the prediction image signal output from the selector 57. Is applied. Similarly, the half pel processing unit 61 in the prediction memory unit 22 shown in FIG. 2 is arranged after the selector 58, and performs a half pel process on the predicted image signal output from the selector 58. Then, the half-pel processing units 62 and 6 in FIG.
3, 64, 65 are deleted. With this configuration, the number of half-pel processing units can be reduced to one third as compared with the prediction memory unit 22 of FIG. Thereby, the cost of the apparatus can be reduced.

In the above-described embodiment, the case where the image is divided into three in the vertical direction and the parallel processing is performed has been described. However, the image may be divided into three in the horizontal direction. It is also possible to divide the image in the horizontal and vertical directions. Further, in the above embodiment, the image is divided into three areas, but the present invention is not limited to this.

[0044]

As described above, according to the encoding apparatus, the encoding method, and the recording medium according to the present invention, for each divided area, image data of another area to be referred to when performing motion compensation. Is stored for each region, so that when an image is divided into a plurality of regions and encoding processing is performed in parallel for each region, the limitation of the motion compensation range can be eliminated, and the motion compensation range can be reduced. Can be expanded.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of an embodiment of an encoding device according to the present invention.

FIG. 2 is a block diagram illustrating a detailed configuration example of a prediction memory unit 22 in FIG. 1;

FIG. 3 is a timing chart for explaining the operation of a prediction memory unit 22;

FIG. 4 is a block diagram illustrating a configuration example of another embodiment of the prediction memory unit 22 in FIG. 1;

[Explanation of symbols]

 11, 21, 31 ME 12, 22, 32 Prediction memory unit 13, 23, 33 DCT 14, 24, 34 Quantizer 15, 25, 35 Dequantizer 16, 26, 36 IDCT 17, 27, 37 Adder 18, 28, 38 Selectors 19, 29, 39 Subtractors 1A, 2A, 3A Coding control unit 51 Upper area F prediction memory 52 F prediction memory 53 Lower area F prediction memory 54 Upper area B prediction memory 55 B prediction memory 56 Lower area B prediction memory 60 to 65 Half pel processing unit 57, 58 Selector 59 Select signal generation unit

Claims (9)

[Claims] 1. An encoding device for dividing an image into a plurality of regions and encoding image data for each region, comprising: a motion vector detecting unit for detecting a motion vector of the image data; First storage means for storing image data of the area; second storage means for storing image data of another area adjacent to the area; and the motion vector detected by the motion vector detection means. And selecting image data of a predetermined area from image data of the area stored in the first storage means and image data of another area stored in the second storage means based on A coding unit for performing motion compensation using image data output from the selection unit. 2. An area corresponding to image data stored in the second storage means is one of upper, lower, left and right with respect to an area corresponding to image data stored in the first storage means. The encoding device according to claim 1, wherein the encoding device is located at at least one of the positions. 3. The memory according to claim 2, wherein the second storage unit stores image data of an area adjacent to an upper side of the area corresponding to the image data stored in the first storage unit, and the first storage unit. 2. The encoding apparatus according to claim 1, further comprising a memory for storing image data of an area adjacent to a lower side of the area corresponding to the image data stored in the means. 4. The memory according to claim 1, wherein the second storage unit stores image data of an area adjacent to the left side of the area corresponding to the image data stored in the first storage unit, and the first storage unit. 2. The encoding apparatus according to claim 1, further comprising: a memory for storing image data of an area adjacent to the right side of the area corresponding to the image data stored in the means. 5. The first storage means and the second storage means comprise a memory for performing forward inter-frame predictive coding and a memory for performing reverse inter-frame predictive coding. The encoding apparatus according to claim 1, wherein the encoding is performed. 6. The image processing apparatus according to claim 1, further comprising a half-pel unit configured to perform a half-pel process on the image data stored in the first storage unit and the second storage unit, The encoding device according to claim 1, wherein the encoding device is provided for each. 7. The encoding apparatus according to claim 1, further comprising a half-pel unit that performs a half-pel process on the image data output from the selection unit. 8. An encoding method for dividing an image into a plurality of regions, encoding image data for each region, detecting a motion vector of the image data, and for each region, In addition to storing the image data, storing the image data of another area adjacent to the area, based on the detected motion vector, the stored image data of the area and the image data of the other area. Encoding means for selectively outputting image data of a predetermined area of the image data and performing motion compensation using the output image data. 9. A recording medium on which a program capable of executing the encoding method according to claim 8 is recorded.