JP2906865B2 - Motion compensation coding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motion compensation coding apparatus for coding a video signal by performing motion compensation.

[0002]

2. Description of the Related Art Generally, a motion compensation coding apparatus compresses a video signal by obtaining a motion vector between frames (or between fields) and performing motion compensation. Among them, the one most frequently used is one that uses inter-frame prediction for obtaining a motion vector between frames, and includes a macro block or a DCT block (hereinafter simply referred to as a block) of the current frame for which motion compensation is performed. In the interpolated image formed from the front or the rear or the front and the rear in terms of time, a position near the block of the current frame where the difference is the smallest is searched for, and the difference from that is encoded. The motion vector obtained as a result is
Predictive encoding represented by two components in the vertical direction and using the motion vector of the immediately preceding block as a predicted value is performed, and variable-length coding is performed on the horizontal prediction error and the vertical prediction error. For an encoding method of an encoding device using such inter-frame prediction, see, for example, “International Standard for Multimedia Encoding”: Maruzen, pp. 114 and 128.
Page 129 to page 129.

[0003]

However, according to the conventional method, when a motion vector detection error occurs, motion compensation is performed using an erroneous motion vector. Because the prediction error of the variable length encoding, the error is propagated when decoding the motion vector, and because the code was also assigned to the motion vector of the same value or a similar value, encoding, There is a problem that the coding efficiency is reduced.

SUMMARY OF THE INVENTION The present invention solves such a drawback of the conventional encoding apparatus, and prevents motion compensation using an erroneous motion vector as compared with the conventional one, or prevents error propagation of a motion vector. , Or to provide a motion compensation coding apparatus that improves the coding efficiency of a motion vector.

[0005]

In order to achieve this object, a motion compensation coding apparatus according to the present invention divides an input television signal into a plurality of small blocks, and obtains a motion vector for each of the small blocks. Block vector detection means,
A large block vector detecting means for dividing an input television signal into large blocks each having a block size equal to the plurality of small blocks, and obtaining a motion vector for each of the large blocks; and each motion vector obtained for each of the large blocks In contrast, a limited range setting unit that sets a limited range according to the motion vector of each of the large blocks, and when each motion vector of the small block in each of the large blocks is within the limited range, A motion compensation switching means for performing motion compensation on small blocks existing within the limited range using the motion vectors of the large blocks and coding the small blocks.

A threshold setting means for setting a threshold value for the number of motion vectors of the small blocks in each of the large blocks existing in the limited range, and a motion vector for each of the large blocks when the motion vector exceeds the threshold value. Motion compensation means for compensating,
Or, the motion vector of the small block present in the limited range, the motion vector encoding means for encoding the difference value with the motion vector of each of the large blocks, or performing motion compensation in each of the large blocks, A motion compensation unit is provided for performing motion compensation on each small block outside the limited range using a difference vector from the motion vector of each large block.

[0007]

With this configuration, the encoding apparatus according to the present invention uses the motion vectors of the blocks of two sizes and the limited range or the limited range and the threshold value, so that when a motion vector detection error occurs, an error occurs. It works to prevent performing motion compensation using the motion vector, prevent error propagation at the time of decoding the motion vector, or improve the coding efficiency of the motion vector.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings.

In FIG. 1, 1 is a subtractor for obtaining a difference between input signals, 2 is an orthogonal transformer for performing orthogonal transformation on an input signal,
3 is a quantizer for quantizing the input signal, 4a and 4b
Is an encoder for encoding the input signal, 5 is an inverse quantizer for performing inverse quantization on the input signal, 6 is an inverse orthogonal transformer for performing inverse orthogonal transformation on the input signal, and 7 is a sum of the input signals. Adders, 8a and 8b are block dividers for dividing the input signal into blocks, 9a and 9b are motion vector detectors for obtaining a motion vector from the input signal, 10 is an output determiner for determining an output from the input signal and outputting it, 11 Is an input terminal for inputting a television signal.

[0010] The following describes the relationship between the components and the operation of the coding apparatus including the above components.

A first frame (or a first field, hereinafter referred to as frame 1) of a television signal input from an input terminal 11 is supplied to a subtracter 1 and a motion vector detector 9.
a and the motion vector detector 9b. Since the motion compensation is not performed on the frame 1, the frame 1 is transmitted to the orthogonal transformer 2 without processing. Then frame 1
Are orthogonally transformed by the orthogonal transformer 2, quantized by the quantizer 3, and transmitted to the encoder 4 a and the inverse quantizer 5. The frame 1 transmitted to the encoder 4a is an encoded signal 12
And the frame 1 input to the inverse quantizer 5
Are inversely quantized by an inverse quantizer 5 and inversely orthogonally transformed by an inverse orthogonal transformer 6, passed through an adder 7, and input to a block divider 8 a and a block divider 8 b. As in the case of the subtractor 1, the adder 7 does not perform processing on the frame 1.

The frame 1 transmitted from the adder 7 is divided into small blocks by a block divider 8a, and divided into large blocks of a size of a plurality of small blocks by a block divider 8b. The small blocks divided by the block divider 8a are transmitted to the motion vector detector 9a and the output determiner 10, and the large blocks divided by the block divider 8b are
It is transmitted to the motion vector detector 9b. The motion vector detector 9a detects a motion vector between the input small block and a second frame (hereinafter, frame 2) which is the next frame input from the input terminal 11, and the motion vector detector 9b A motion vector is detected between the input large block and the frame 2, and the motion vector detected by each motion vector detector is transmitted to the output determiner 10.

The output determiner 10 sets a limited range in accordance with each large vector for the motion vector (hereinafter, referred to as a large vector) input from the motion vector detector 9b. Is output as the motion vector 14 of each small block when the motion vector (hereinafter, referred to as a small vector) is within the limited range, and when the motion vector is not within the limited range. . The outputted motion vector 14 is encoded by the encoder 4b and output as an encoded signal 15. The output determiner 10 outputs the motion vector and, at the same time, outputs the video signal 13 of the small block input from the block divider 8a. The outputted small block video signal 13 is transmitted to the subtractor 1 and the adder 7.

The frames after the frame 2 are subtracted by the subtracter 1
Then, the difference from the transmitted small block video signal 13 is obtained and transmitted to the orthogonal transformer 2. Thereafter, adder 7
The processing up to is the same as in the case of frame 1. Adder 7
Then, the sum of the signal transmitted from the inverse orthogonal transformer 6 and the small block transmitted from 19 is obtained and transmitted to the block divider 8a and the block divider 8b. Subsequent processing is the same as in the case of frame 1.

As described above, according to the present embodiment, the input television signal is divided into a plurality of small blocks by the block divider 8a, and the motion vector detector 9a detects the small block vector for obtaining the motion vector of each of the small blocks. Means and a large block vector detecting means for dividing an input television signal into large blocks having a block size corresponding to the plurality of small blocks by a block divider 8b and obtaining a motion vector of each of the large blocks by a block divider 9b. Means, a limited range setting means for setting a limited range according to the motion vector of each large block for each motion vector obtained for each large block by the output determiner 10, and When each motion vector of the small block is within the limited range, the motion vector of each of the large blocks is used, Motion compensation coding means for performing motion compensation on small blocks existing in the frame and coding the same, so that when a large number of similar motion vectors appear in a large block, they are regarded as one vector and the motion vector difference value And improve the coding efficiency of the motion vector.

Next, a second embodiment of the present invention will be described with reference to the drawings. In FIG. 2, 1 is a subtractor, 2
Is an orthogonal transformer, 3 is a quantizer, 4a and 4b are encoders, 5 is an inverse quantizer, 6 is an inverse orthogonal transformer, 7 is an adder,
8a and 8b are block dividers, 9a and 9b are motion vector detectors, 10a is an output determiner, and 11 is an input terminal.

A description will be given of the relationship between the components and the operation of the encoding device including the above components.

Frame 1 input from input terminal 11
Are processed in the same manner as in the first embodiment up to the block divider 8a and the block divider 8b. In this embodiment, the small blocks divided by the block divider 8a and the large blocks detected by the block divider 8b are transmitted to an output determiner 10a, a motion vector detector 9a, and a motion vector detector 9b. The motion vector detector 9a and the motion vector detector 9b detect a motion vector between the frame 1 and the frame 2 as in the first embodiment, and input the detected motion vector to the output determiner 10a.

In the output determiner 10a, a limited range corresponding to each large vector is set for the large vector of each large block, as in the first embodiment. A threshold is set for the number of small vectors of the small block to be executed. If the number of small vectors exceeds the threshold, the output determiner 10a outputs the video signal 16 and the large vector 17 of the large block input from the block divider 8b. The video signal 16 and the small vector 17 of the input small block are output. The output video signal of the large block or small block 16 is processed in the same manner as in the first embodiment, and the output motion vector 17 is encoded by the encoder 4b.
It is output as an encoded signal 18. Subsequent processing is the first
This is the same as the embodiment or the case of the frame 1.

As described above, in the present embodiment, the input television signal is divided into a plurality of small blocks by the block divider 8a, and the small block motion vector for obtaining the motion vector of each of the small blocks by the motion vector detector 9a. Detecting means and a large block in which the input television signal is divided into large blocks whose block size is equal to the plurality of small blocks by a block divider 8b, and a motion vector is obtained by each of the large blocks by a motion vector detector 9b. A motion vector detecting means, an output determiner 10a, a limited range setting means for setting a limited range corresponding to each large vector for a motion vector of each large block, and a small vector number of small blocks existing within the limited range. A threshold setting means for setting a threshold value with respect to, and an output determiner 1 when the number of small vectors exceeds the threshold value.
0a is provided with a motion compensation means for outputting the large block video signal 16 and the large vector 17 inputted from the block divider 8b, so that when a small block erroneously detects a motion vector, motion compensation is performed with an erroneous motion vector. Is performed, and the motion vector becomes one motion vector of the large block, so that the coding efficiency of the motion vector is also improved.

Next, a third embodiment of the present invention will be described with reference to the drawings. In FIG. 3, 1 and 1a are subtractors, 2 is an orthogonal transformer, 3 is a quantizer, 4a and 4b are encoders, 5 is an inverse quantizer, 6 is an inverse orthogonal transformer, 7 is an adder, 8a and 8b are block dividers, 9a and 9b are motion vector detectors, 10b is an output determiner, and 11 is an input terminal.

A description will be given of the connection and operation of the encoding device having the above components with each other.

Frame 1 input from input terminal 11
Are the motion vector detector 9a and the motion vector detector 9
The processing up to b is performed in the same manner as in the first embodiment. In the present embodiment, the small vector detected by the motion vector detector 9a and the large vector detected by the motion vector detector 9b are:
The output is transmitted to the output determiner 10b and the subtractor 1a. Subtractor 1
In a, a difference between a large vector and a small vector (hereinafter, referred to as a difference vector) is obtained and transmitted to the determiner 10b.
In the determiner 10b, a limited range corresponding to each large vector is set for the large vector of each large block as in the case of the second embodiment, and the number of small vectors of the small blocks existing within the limited range is set. To set the threshold. Output determiner 10b
Outputs the video signal 19 of the small block, outputs the difference vector 20 input from the subtractor 1a when the number of small vectors exceeds the threshold, and outputs the small vector 20 when the number is smaller than the threshold. . Output small block video signal 1
9 is processed in the same manner as in the first embodiment, and the output difference vector or small vector 20 is encoded by the encoder 4b and output as an encoded signal 21. Subsequent processing is the same as in the first embodiment or the case of frame 1.

As described above, according to this embodiment, the input television signal is divided into a plurality of small blocks by the block divider 8a, and the small block motion vector for obtaining the motion vector of each small block by the motion vector detector 9a. Detecting means and a large block in which the input television signal is divided into large blocks whose block size is equal to the plurality of small blocks by a block divider 8b, and a motion vector is obtained by each of the large blocks by a motion vector detector 9b. A motion vector detecting means, an output determiner 10b for setting a limited range corresponding to each large vector for a motion vector of each large block, and a small vector number of small blocks existing within the limited range. Threshold setting means for setting a threshold for the small block, and when the number of small vectors exceeds the threshold, And a motion vector encoding unit that encodes the motion vector as a difference vector by the encoder 4b. Since the blocks within the limited range transmit the difference vector, the motion vector coding efficiency is improved, Even when the transmitted difference vector is decoded, the propagation of the error can be prevented since each of the small vectors obtains the difference value from the large vector.

Next, a fourth embodiment of the present invention will be described with reference to the drawings. In FIG. 4, 1 and 1a are subtractors, 2 is an orthogonal transformer, 3 is a quantizer, 4a and 4b are encoders, 5 is an inverse quantizer, 6 is an inverse orthogonal transformer, 7 is an adder, 8a and 8b are block dividers, 9a and 9b are motion vector detectors, 10c is an output determiner, and 12 is an input terminal.

The relationship between the components and the operation of the encoding device having the above components will be described below.

Frame 1 input from input terminal 11
Are the motion vector detector 9a and the motion vector detector 9
The processing up to b is performed in the same manner as in the second embodiment. In the motion vector detector 9a and the motion vector detector 9b, the third
The small vector and the large vector are output to the output determiner 10c and the subtractor 1a in the same manner as in the embodiment.
In the subtractor 1a, a difference vector is obtained in the same manner as in the third embodiment, and is transmitted to the determiner 10c. Judgment device 10c
In the same manner as in the second embodiment, a limited range corresponding to each large vector is set for a large vector of each large block, and a threshold is set for the number of small vectors of small blocks existing within the limited range. Set. When the number of small vectors exceeds the threshold value, the output determiner 10c
The difference vector 23 of the small block outside the limited range is output from the large vector inputted from the small vector and the difference vector inputted from the subtractor 1a. The video signal 22 of the outer small block is output. The output video signal 22 is processed in the same manner as in the first embodiment, and the output motion vector 23 is coded by the coder 4b.
4 is output. Subsequent processing is the same as in the first embodiment or the case of frame 1.

As described above, in the present embodiment, the input television signal is divided into a plurality of small blocks by the block divider 8a, and the small block motion vector for obtaining the motion vector of each of the small blocks by the motion vector detector 9a. Detecting means and a large block in which the input television signal is divided into large blocks whose block size is equal to the plurality of small blocks by a block divider 8b, and a motion vector is obtained by each of the large blocks by a motion vector detector 9b. A motion vector detecting means, an output determiner 10b for setting a limited range corresponding to each large vector for a motion vector of each large block, and a small vector number of small blocks existing within the limited range. And a threshold setting means for setting a threshold value for the motion vector detector 9b when the number of small vectors exceeds the threshold value. A large vector, a subtractor 1a
Out of the input difference vectors, the difference vector 23 of the small block outside the limited range is output.
By outputting the video signal of the large block input from b and the video signal 22 of the small block outside the limited range, motion compensation is performed in the large block, and the small block outside the limited range is calculated using the difference vector of the large block. With the provision of the motion compensation means for performing motion compensation, it is possible to reduce the code amount of the motion vector and perform efficient encoding.

In the second embodiment, the large vector is coded when the number of small blocks within the limited range exceeds the threshold, but the average value of the small vectors may be used.

[0030]

As described above, according to the motion compensation coding apparatus of the present invention, a small block vector detecting means for dividing an input television signal into a plurality of small blocks and obtaining a motion vector in each of the small blocks. A large block vector detecting means for forming a large block from a plurality of the small blocks, and obtaining a motion vector in each of the large blocks;
For each motion vector obtained in each of the large blocks, a limited range setting unit that sets a limited range according to the motion vector of each of the large blocks, and each motion vector of the small block in each of the large blocks, When present in the limited range, using the motion vector of each large block,
By providing a motion compensation switching means for performing motion compensation on small blocks existing in the limited range and coding the same, when a large number of similar motion vectors appear in a large block, they are regarded as one vector. It is possible to reduce the difference value of the motion vector and improve the coding efficiency of the motion vector.

A threshold setting means for setting a threshold for the number of motion vectors of the small blocks in each of the large blocks existing in the limited range; and a motion vector for each of the large blocks when the motion vector exceeds the threshold. With the configuration including the motion compensating means for compensating, when a small block erroneously detects a motion vector, it is possible to prevent performing motion compensation with an erroneous motion vector, and the motion vector becomes one motion vector of a large block. Therefore, the coding efficiency of the motion vector is also improved.

The motion vector coding means for coding a motion vector of a small block existing within the limited range by calculating a difference value between the motion vector of each large block and the motion vector of the large block is provided. Since the blocks within the limited range transmit the difference vector, the coding efficiency of the motion vector is improved, and when decoding the transmitted difference vector, each of the small vectors obtains a difference value from the large vector. Error propagation can be prevented.

A motion compensation unit for performing motion compensation on each of the large blocks and performing motion compensation on each of the small blocks outside the limited range using a difference vector from the motion vector of each of the large blocks. By doing so, it is possible to reduce the code amount of the motion vector and perform efficient coding.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a motion compensation coding apparatus according to a first embodiment of the present invention.

FIG. 2 is a configuration diagram of a motion compensation coding apparatus according to a second embodiment of the present invention.

FIG. 3 is a configuration diagram of a motion compensation coding apparatus according to a third embodiment of the present invention.

FIG. 4 is a configuration diagram of a motion compensation coding apparatus according to a fourth embodiment of the present invention.

[Explanation of symbols]

 1, 1a Subtractor 2 Orthogonal transformer 3 Quantizer 4a, 4b Encoder 5 Inverse quantizer 6 Inverse orthogonal transformer 7 Adder 8a, 8b Block divider 9a, 9b Motion vector detector 10, 10a, 10b , 10c Output determiner 11 Input terminal

Claims (7)

(57) [Claims] A motion compensation coding apparatus for performing motion compensation by coding an input television signal using a motion vector indicating a motion amount and a direction of an image between frames (or between fields). Divide the John signal into multiple small blocks,
A small block vector detecting means for obtaining a motion vector for each of the small blocks; and a large block for dividing an input television signal into large blocks having a block size corresponding to the plurality of small blocks, and obtaining a motion vector for each of the large blocks. Block vector detecting means, a limited range setting means for setting a limited range according to the motion vector of each of the large blocks for the motion vector obtained in each of the large blocks, When each motion vector is within the limited range, using a motion vector of each of the large blocks, motion compensation switching coding means for performing motion compensation on a small block existing within the limited range and coding. A motion compensation coding apparatus comprising: 2. A motion compensation encoding apparatus for performing motion compensation by coding an input television signal using a motion vector indicating a motion amount and a direction of an image between frames (or between fields). Divide the John signal into multiple small blocks,
A small block vector detecting means for obtaining a motion vector for each of the small blocks; and a large block for dividing an input television signal into large blocks having a block size corresponding to the plurality of small blocks, and obtaining a motion vector for each of the large blocks. Block vector detecting means; limiting range setting means for setting a limiting range according to the motion vector of each large block with respect to the motion vector obtained in each of the large blocks; Threshold value setting means for setting a threshold value for the number of motion vectors of small blocks in the block; and motion compensation means for performing motion compensation in each of the large blocks when the number of motion vectors exceeds the threshold value. Motion compensation coding apparatus. 3. A motion compensation coding apparatus for performing motion compensation by coding an input television signal using a motion vector indicating a motion amount and a direction of an image between frames (or between fields). Divide the John signal into multiple small blocks,
A small block vector detecting means for obtaining a motion vector for each of the small blocks; and a large block for dividing an input television signal into large blocks having a block size corresponding to the plurality of small blocks, and obtaining a motion vector for each of the large blocks. Block vector detecting means; limiting range setting means for setting a limiting range according to the motion vector of each large block with respect to the motion vector obtained in each of the large blocks; Threshold setting means for setting a threshold for the number of motion vectors of the small blocks in the block; and, when the number of motion vectors exceeds the threshold, the motion vector of the small block existing in the limited range, Motion vector coding means for obtaining and coding a difference value between the motion vector and the motion vector apparatus. 4. A motion compensation coding apparatus for performing motion compensation by coding an input television signal using a motion vector indicating a motion amount and a direction of an image between frames (or between fields). Divide the John signal into multiple small blocks,
A small block vector detecting means for obtaining a motion vector for each of the small blocks; and a large block for dividing an input television signal into large blocks having a block size corresponding to the plurality of small blocks, and obtaining a motion vector for each of the large blocks. Block vector detecting means; limiting range setting means for setting a limiting range according to the motion vector of each large block with respect to the motion vector obtained in each of the large blocks; Threshold setting means for setting a threshold for the number of motion vectors of the small blocks in the block; and when the number of motion vectors exceeds the threshold, perform motion compensation in each of the large blocks, and each small block outside the limited range. Motion compensation means for performing motion compensation using a difference vector from the motion vector of each large block. Compensation encoding device. 5. The motion compensation coding apparatus according to claim 1, wherein the limited range is a magnitude of a motion vector. 6. The motion compensation coding apparatus according to claim 1, wherein the limited range is a direction of a motion vector. 7. The motion compensation coding apparatus according to claim 1, wherein the limited range is a magnitude and a direction of a motion vector.