JP2002165221A - Dynamic image data compression system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a dynamic image data compression system, in which a prescribed data compression efficiency can always be maintained, irrespective of the data content of dynamic image data to be input. SOLUTION: Difference block region data(r) and difference block image data(d), which are created by a difference block extractor 10, are input to a difference block coupler 12. From among the difference block region data(r), the data which expresses mutually adjacent blocks are selected, and the adjacent blocks are coupled to each other, so as to be replaced with difference block region data in one unit; and image data which correspond to the adjacent difference block region data are coupled to each other, so as to be replaced with coupled difference block image data in one unit.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for compressing moving image data comprising a plurality of frames, and more particularly to a method for dividing each frame into a plurality of blocks and describing a difference between the frames for each block. The video data compression method.

[0002]

2. Description of the Related Art Generally, image data has a large amount of information. However, in the case of moving images, in particular, the image data is remarkably composed of a plurality of frame strings, and requires a large-capacity storage device for recording. Therefore, conventionally, the amount of information is suppressed by a data compression technique so that the storage area of the storage device is reduced. However, in the case of moving image data, the data compression in each frame constituting an image is performed. In addition, it is possible to obtain data compression while suppressing redundancy by utilizing the correlation between frames.

In this case, a frame is divided into a plurality of blocks, and data describing an area occupied by the block in the frame is output to a table only for a block whose data has changed in comparison with the immediately preceding frame. Conventionally, a data compression method based on a so-called inter-block difference method of removing redundancy between frames by outputting image data in the block to a storage device has been conventionally known. It has been disclosed.

[0004] Hashihara, “Blocked Error Diffusion Me
thod for Image Sequences ", IBM TDB Vol.33, No.10A, pp.94-95, March 1991.
It discloses a method in which a plurality of blocks divided from a frame are sequentially compared with the blocks of the immediately preceding frame, and only data of a block in which a change equal to or more than a predetermined reference is detected is stored.

FIG. 7 is a diagram showing a difference when moving image data composed of a series of a plurality of frames numbered in order from 0 is compressed between frames according to a method represented by the above-mentioned document. FIG. 9 illustrates a state in which block area data and image data are output to a storage device.

In this example, first, each frame is divided into i in the horizontal direction and j in the vertical direction, and divided into (i × j) blocks. Each block is assigned a block number "b" in order from the upper left block to the lower right block. In this example, since the number of blocks i in the horizontal direction is i = 4 and the number of blocks in the vertical direction j = 4, the number of blocks is 16, and b = 0, 1,.

Next, an arbitrary number f (f
= 0 to a predetermined number), the image data of a block of a certain number b in the frame of number b is defined as d (f, b), and this is set to the block of the block of number b in the frame of number f−1 immediately before the frame of number f. It is compared with the image data d (f-1, b) to determine whether or not the image data has changed.

Then, the block number b of the block in which the change has been detected is stored as difference image block area data r on the table of the frame number f on the storage device, and the image data d (f, b) of the block is stored. The difference block image data d is stored on the page of the frame number f on the storage device, and this is sequentially repeated for each block and each frame.

Incidentally, FIG. 7 shows that among the blocks constituting the frame of the number f, the block number b is 3,
The state in which the above-mentioned change occurs in blocks 6, 7, 12, and 14, and the corresponding difference block area data r and difference block image data d are stored in a table and a page, respectively. Here, the difference block area data r may have any format as long as the block number can be uniquely specified.

On the other hand, in Japanese Patent Laid-Open Publication No. Hei 4-368086, one of the binary numbers that can be represented by one bit is described at a position determined by the block number b instead of the description method of the difference block area data r. By using a so-called bitmap method, one of the above binary numbers is described at a position corresponding to a certain block in a certain frame. In this case, a method for describing the other of the above binary numbers is disclosed.

[0011]

The prior art described above does not give sufficient consideration to improving the compression efficiency of moving image data, and has a problem that the compression ratio of moving image data remains unsatisfactory. Was. First, among the conventional techniques, the data compression method described with reference to FIG. 7 can improve the compression ratio by suppressing redundancy between frames, but requires the block area data separately from the image data itself. May be accompanied by a decrease in efficiency.

In this method, since the difference block area data amount is proportional to the number of changed blocks, the difference block area data amount increases as the number of changed blocks increases.

This means that if the number of changed blocks becomes larger than a certain number, a larger amount of data is required than if the image data of the frame is simply output as it is, and the compression is performed. Meaning is lost.

On the other hand, in the case of the prior art disclosed in the above-mentioned publication, the amount of difference block area data depends only on the number of pixels in the block, regardless of the number of changed blocks. Although the data compression efficiency as a whole does not deteriorate even if the number of blocks increases, it is necessary to always output the difference block area data having a constant data amount to the storage device regardless of the number of changed blocks.

For example, (i × m) pixels in the horizontal direction, and
In the case of a frame of (j × n) pixels, even if there is no changed block, (i × m × j × n) bits of data are required, and the data amount per frame is (i × m × j ×
It cannot be smaller than n) bits.

For this reason, there is a problem that the compression efficiency cannot be increased even when a frame with a small change continues, and a certain lower limit is given to the data amount of the entire moving image data.

An object of the present invention is to provide a moving image data compression system which can always maintain a predetermined data compression efficiency irrespective of the data content of the input moving image data.

[0018]

SUMMARY OF THE INVENTION It is an object of the present invention to divide each frame of moving image data composed of a plurality of temporally continuous frames into a plurality of blocks, and to divide the frames into frames in order. Is compared with the data of the corresponding block of the immediately preceding frame, and for each of the blocks whose data has changed, the image data corresponding to the block is held as differential block image data, and the block corresponding to the block is stored. In the moving image data compression method of holding a position of a difference block in a frame as difference block region data, in the held difference block region data, difference block region data located at positions adjacent to each other in the frame in the held difference block region data Means for linking each other and replacing and holding one unit of differential block area data; Means for holding the image data corresponding to the difference block area data as difference block image data, and using the difference block image data after replacement and the difference block image data held by these means, Is achieved in such a way that

As a result, when the changed blocks appear adjacently on the screen, those blocks are concatenated and processed as one unit of data.
Data volume can be reduced.

On the other hand, if the number of changed blocks increases, the probability that they appear adjacently increases. Therefore, even if the number of changed blocks increases, the increase in the amount of differential block area data is relatively small. Therefore, even if the data content of the input moving image data changes, it is possible to easily maintain a predetermined compression ratio.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a moving picture data compression system according to the present invention will be described in detail with reference to an embodiment. FIG. 1 shows a configuration of an entire system according to an embodiment of the present invention. Here, a differential block extractor 10 and a differential block connector 12 constitute an encoding side for compressing data, and a block image. The data retrieving unit 14 and the frame composing unit 16 constitute a decoding side for decompressing the data and returning it to the original image data, and the storage device managing unit 18 and the storage device 20 constitute a common part.

The moving image data to be subjected to data compression is input to the difference block extractor 10. At this time, the frame having the number f to be coded and the immediately preceding number f f-1 frames are paired,
The moving image data which is sequentially input and returned by the frame constructor 16 is output from the output buffer 26.
It is supplied to.

First, the encoding side will be described. First,
The difference block extractor 10 corresponds to each of the blocks constituting the corresponding frame in the moving image data input for each frame, in each of the frames except the first frame, in the frame immediately before the frame. Compare with block.

When there is a change in pixel data between blocks, difference block area data r representing the position of the block in the frame and difference block image data d corresponding to the block are output.

Therefore, the processing up to this point is the same as the inter-block difference method according to the prior art described with reference to FIG. 7, but in this embodiment, next, the difference block area data r and the difference block image data d Is further supplied to the difference block coupler 12.

Therefore, the differential block coupler 12 first determines whether or not there is any adjacent data in the differential block area data r. The block area data are connected to each other and replaced with one unit of data, which is referred to as connected difference block area data.

The meaning of the adjacent at this time is that the block of the data r has an adjacent block on the frame plane in FIG. 7, and the adjacent directions at this time are the horizontal direction and the vertical direction. It may be in any direction, and includes a case where two or more blocks are arranged.

This will be described with reference to FIG. 7. For example, as shown in the drawing, if there is the difference block area data r of the blocks 5, 6, and 7 in this frame in the table of number f, these difference blocks Area data r is adjacent
When the difference block region data r of the blocks 0 and 4 are present, these difference block region data r are also adjacent (vertical direction).
The obtained difference block area data are obtained.

Then, for each of the linked difference block area data, the difference block area data and the difference block image data corresponding to the difference block area data are output and supplied to the storage device manager 18.

Hereinafter, the frame reconstructing process performed as a part of the process of reproducing a frame will be described. However, the purpose of frame reconstruction is not limited to frame playback,
It may be performed as preprocessing for extracting a statistic from a frame image.

On the other hand, on the decoding side, the block image data searcher 14 inputs the number f 'of the frame to be reconstructed from the frame reproducer 13 and reconstructs each of the blocks constituting the frame from the first frame. In the range up to the target frame, the storage device 2 including the number b of the block and the last differential block area data including the block
The table numbers in 0 are paired, a list of them is constructed, and the list is output to the frame constructor 16.

The frame constructor 16 stores, for each of the blocks specified by the elements in the list, the storage device 20 storing the image data of the block based on the input list of the block number and the table number. And the address within the page are calculated.

Then, the page of the storage device 20 is accessed using the calculated page number and address, the image data of the block is fetched, and each of the obtained image data is combined to form a frame to be reconstructed.

At this time, the conversion process from the page number and the address to the physical address on the storage device 20 is performed by the storage device manager 18, but a known method can be used for the conversion to the physical address. Then, the description is omitted.

Next, the differential block extractor 10 on the encoding side
And a data compression process by the difference block coupler 12, a buffer storage format of the difference block area data and a buffer storage format of the image data created by the difference block coupler 12, and a storage device manager 18 which is a common part thereof. Details of the processing by the device 20 will be sequentially described.

<Difference Block Extractor 10> FIG. 2 shows details of the difference block extractor 10, and
Stores the image data of the frame of the number f for which the difference block is to be extracted, and the reference image buffer 52 stores
The image data of the frame of the number f-1 from which the difference block is extracted is stored.

From these buffers 50 and 52,
Corresponding to each of the blocks constituting the frame of the number f, the input image data d (f, b) of the block and the reference image data d (f-1, b) of the block are sequentially read out and the difference calculator The difference between the image data is sequentially calculated, and at the same time, from the block area data constructor 54, corresponding to each of the blocks constituting the frame of the number f, Block numbers b are sequentially output.

The output of the difference calculator 56 is input to a comparator 58, where it is determined whether there is a difference. When there is a difference, the gates 60 and 62 are opened, the image data d (f, b) is supplied to the output buffer 64 as the difference block image data d, and the block number b of the corresponding block is set in the difference block area. The data is supplied to the output buffer 66 as data r.

When the above procedure has been executed for all the blocks of the frame of the number f, the contents of the input image buffer 50 are transferred to the reference image buffer 52, and
The image data of the next frame is written into the input image buffer 50. Then, the above-described difference block extraction processing is executed for all frames in addition to the processing by the difference block coupler 16 described later.

As described above, the processing until the difference block image data d is stored in the output buffer 64 and the difference block area data r is stored in the output buffer 66 by the difference block extractor 10 is shown in FIG. This is the same as the inter-block difference method according to the prior art described in FIG.

<Difference Block Linker 12> FIG. 3 shows the details of the difference block coupler 12 and the difference block extractor 1
0, the difference block area data r and the difference block image data d are input for each frame, and the following processes (a), (b), and (c) are executed in accordance with the procedures to be described later. It is.

(A) From the input difference block area data r, the data representing the blocks adjacent to each other is selected. (b) The adjacent difference block area data is connected to each other and replaced with one unit of difference block area data. (c) The image data corresponding to the adjacent difference block area data is connected to each other and replaced with one unit of connected difference block image data.

Next, the above procedures 1 to 5 will be described. Procedure All of the difference block area data r created by the difference block extractor 10 is input to the input difference block area data buffer 7.
In addition to the input to 0, all of the differential block image data d is input to the input differential block image data buffer 72.

Procedure From the input differential block area data buffer 70, the differential block area data r is extracted one unit at a time, selected as the link target block area data 76, and selected from the input differential block image data buffer 72. Similarly, the difference block image data d to be extracted is extracted one unit at a time, and is selected as the connection target block image data 92.

Procedure From the input differential block area data buffer 70, differential block area data r different from the link target block area data 76 is selected, and this is used as the link candidate block area data 74. From the image data buffer 72, select difference block image data d corresponding to the connection candidate block area data,
This is referred to as connection candidate block image data 90.

Procedure The link target block area data 76 and the link candidate block area data 74 are input to the horizontal adjacency determiner 78, and it is determined whether the block area data can be connected to each other horizontally. The determination at this time is made based on whether or not the data blocks are adjacent in the horizontal direction on the frame plane.

Procedure In the above procedure, if it is determined that horizontal connection is possible, 2
Gates 82 and 84 are opened, and the link candidate block area data 74 and link target block area data 76 are input to a block area data link 88, where the block area data linked in one unit, that is, the linked difference The block area data r 'is output and replaced with the link target block area data 76.

When it is determined that horizontal connection is possible, two gates 96 and 98 are also opened, and the connection candidate block image data 90 and the connection target block image data 92 are converted to block image connectors. The block image data is also input to the block 102, and the block image data is also linked in one unit, that is, the linked difference block image data d ′ is output and replaced with the link target image data 92.

On the other hand, when it is determined in the above procedure that the horizontal connection is impossible, the gate 80 is opened, so that the connection candidate block area data 74 is additionally written in the unconnected block area data buffer 86 and the gate 94 is also opened. The connected candidate block image data 90 is added to the unconnected block image data buffer 100.

Procedure It is checked whether or not block area data that has not been selected as the connection candidate block area data 74 remains in the input differential block area data buffer 70. Is selected and replaced with the link candidate block area data 74.

At the same time, block image data corresponding to the connected candidate block area data after replacement is selected from the input differential block image data buffer 72, replaced with the connected candidate block image data 90, and the procedure returns.

On the other hand, if there is no block area data not selected as the connection candidate block area data 74 in the input difference block area data buffer 70, the link target block area data 76 is replaced with the connected difference block area data. The content of the input difference block region data buffer 70 is replaced with the content of the unconnected block region data buffer 86 as the region data r 'is additionally written in the connected block region data buffer 104.

Further, the block image data 92 to be connected is additionally written as connected differential block image data d 'in the connected block image data buffer 106, and the contents of the input differential block image data buffer 72 are stored in the unconnected block image data buffer 100. Buffer 8
After clearing the buffer 6 and the buffer 100, the procedure returns to the procedure.

Procedure After the horizontal connection processing is completed for all the block area data in the input differential block area data buffer 70 according to the above-described procedures, the input differential block area data buffer 70 is connected to the connected block area data buffer 1.
After copying all of the horizontally connected block area data in 04 and clearing the buffer 104, vertical connection processing is executed this time in the same manner as in the above procedures.

The processing at this time is as follows except that switches described above and below the horizontal connection determiner 78 are switched, and a vertical connection determiner 79 is used in place of the horizontal connection determiner 78. Since the procedure in the horizontal connection process described above is the same as the process according to the procedure, the detailed description is omitted.

The determination as to whether or not the vertical connection is possible at this time is also made based on whether or not there is a block adjacent to the data block on the frame plane, but this time in the vertical direction. Check the neighbor status.

Procedures The above-described procedures are executed for the horizontal connection processing and the vertical connection processing. When the connected block area data and the connected block image data are created,
4 and 106 to the storage device manager 18 to end the processing.

<Differential Block Area Data Buffer Storage Format> Differential block extractor 10 and differential block coupler 1
The buffer storage format of the differential block area data created by 2 is a format that describes two pairs of binary numbers with L as the number of digits defined by the following equation.

L≡ [log ₂ (i × j)] Here, when the actual number given by [log ₂ (i × j)] on the right side is z, the number of digits L is greater than or equal to the value of z. It shall be the smallest integer.

Then, as shown in FIG. 4, one L-digit binary number describes the number of the block located at the upper left of the differential block area, and the other L-digit binary number describes the lower right corner of the differential block area. The difference block area is designated by describing the number of the block located.

FIG. 4 shows a case where the number of differential block area data is R (0 to R-1), and each of the differential block area data is located at the upper left of the differential block area. The data is stored in the buffer in ascending order of the block number.

<Buffer storage format of difference block image data> Difference block extractor 10 and difference block coupler 1
As shown in FIG. 5, the buffer storage format of the differential block image data created by Step 2 is a format in which the image data of each block forming the block area is arranged in order from the upper left to the lower right of the block area.

Here, FIG. 5 shows that the number of the block located at the upper left of the block area is p, and the number of the block located at the lower right is p + i × (H−1) + W−1. This is a case of differential block image data corresponding to a block area formed by connecting W × H blocks.

The storage format of the image data of one block of the differential block image data is as shown in FIG.
This is a format in which pixel data is arranged in order from the upper left to the lower right of the block.

At this time, each of the differential block image data is stored in the buffer in the same order as the corresponding differential block area data, so that the block area and the image data can always be associated one-to-one. I have.

At this time, as is apparent from FIG. 5, since the difference block image data is sequentially stored in units of one block, if a block number is designated, the storage address of the image data in that block is changed. Determined uniquely.

Therefore, if the block number, the table number for storing the area data of the block and the page number for storing the image data of the search block are determined, the address on the page for storing the image data of the block is uniquely determined. It is determined.

<Storage device manager 18> Storage device manager 1
8 inputs the number f of the frame to be encoded, inputs the connected difference block region data r 'from the buffer 104 of the difference block connector 12, and inputs the connected difference block image data d' from the buffer 106. .

At this time, the table number t of the difference block area data storage buffer 22 in the storage device 20 and the page number p of the difference block image data storage buffer 24 are selected corresponding to the frame number f.

The difference block area data r and the difference block image data d input from the storage device manager 18 to the storage device 20 are stored in the selected table of number t and page of number p, respectively. The storage format of the difference block area data on each table and the difference block image data on each page is as described above in each section.

Next, the processing on the decoding side by the block image data retrieving unit 14 and the frame composing unit 16 will be described in detail. Needless to say, the processing on the decoding side is executed after the processing on the encoding side is completed.

<Block Image Data Search Unit 14> The block image data search unit 14 inputs the number f ′ of the frame to be reconstructed, and the table number t ′ of the corresponding difference block area data buffer and the difference block image Select the page number p 'of the data buffer.

For each of the blocks constituting the frame, a table for storing the difference block area data including the block is searched from the table with the number t 'to the table 0. Of the image data of the block in the page corresponding to the.

Then, a list including a set of a page number (which matches the table number), a block number, and the address in the page corresponding to the searched table number is created and output to the frame constructor 16.

<Frame Configurator> Frame Configurator 16
Is specified by the page number of the set for each of the block numbers of the set in the list, based on a list formed by the block image data search unit 14 and configured by the set of the block number, the page number, and the address in the page. Page and
The address specified by the address within the page of the set is accessed to acquire the image data of the corresponding block.

Then, each of the acquired difference block image data is copied to an address of the output buffer 26 corresponding to the block number, thereby completing the reconstruction of the frame to be reconstructed.

As a result, according to this embodiment, when changed blocks appear adjacent on the screen, those blocks are concatenated and processed as one unit of data. Thus, a reduction in the amount of data can be obtained.

Here, as the number of changed blocks increases, the probability that they appear adjacently increases. Therefore, according to this embodiment, even if the number of changed blocks increases, the increase in the amount of difference block area data is relatively small, and the data content of the input moving image data changes. However, the predetermined compression ratio can be easily maintained.

[0079]

According to the present invention, even if the data content of the input moving image data changes, there is no possibility that the amount of data in the block area will increase. Can be obtained sufficiently.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram showing an embodiment of a moving image data compression system according to the present invention.

FIG. 2 is a system configuration diagram showing details of a differential block extractor according to an embodiment of the present invention.

FIG. 3 is a system configuration diagram showing details of a differential block coupler according to an embodiment of the present invention.

FIG. 4 is an explanatory diagram showing a storage format of differential block area data according to an embodiment of the present invention.

FIG. 5 is an explanatory diagram showing a storage format of differential block image data according to an embodiment of the present invention.

FIG. 6 is an explanatory diagram showing a storage format of one block of image data constituting differential block image data according to an embodiment of the present invention.

FIG. 7 is an explanatory diagram showing a storage format of difference block area data and difference block image data according to the inter-block difference method.

[Explanation of symbols]

 Reference Signs List 10 difference block extractor 12 difference block coupler 13 frame reproducer 14 block image data searcher 16 frame constructor 18 storage device manager 20 storage device 22 difference block region data buffer 24 difference block image data buffer 26 output buffer

Claims (1)

[Claims] 1. Each frame of moving image data composed of a plurality of temporally continuous frames is divided into a plurality of blocks, and the correspondence between the blocks of the frame and the immediately preceding frame is sequentially determined for each frame. Compare the data of the blocks to be changed, and for each of the blocks whose data has changed,
In the moving image data compression method of holding the image data corresponding to the block as difference block image data and holding the position of the difference block corresponding to the block in the frame as difference block area data, Means for connecting the differential block area data at positions adjacent to each other in the frame in the differential block area data and replacing and storing the differential block area data as one unit of differential block area data; Means for holding image data corresponding to the image data as difference block image data, and the moving image data is represented by the difference block area data after replacement held by these means and the difference block image data. A moving image data compression method, characterized in that: