JP2003339047A - Image compression device, image decompression device, image compression/decompression device, image compression method, image decompression method, program, and recording medium recording the program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide image compression/decompression techniques for compressing an image in matching with requirements of a user by designating a region of interest of the image according to the user of the image, setting a quantization rate different from that for other regions to the designated region of interest, and controlling the compression on the basis of the set quantization rate. <P>SOLUTION: An image compression/decompression device 20 includes: an image division unit 21 for dividing an image into a plurality of regions; a region designation unit 24 for designating one region of interest or more; a quantization rate determination unit 23 for setting a quantization rate to each divided region; and a compression control unit 22 for controlling the compression for each region on the basis of the set quantization rate. The image division unit 21 divides the image into a plurality of regions, the region designation unit 24 designates the region of interest in the image, the quantization rate determination unit 23 sets the quantization rate different from that of other regions to the determined region of interest when setting the quantization rate to each division region, and the compression control unit 22 applied compression control to each region on the basis of the set quantization rate. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image compression apparatus, an image decompression apparatus, an image compression / decompression apparatus, an image compression method, an image decompression method, a program, and a recording medium recording the program, and more specifically, an image An image compression device, an image decompression device, an image compression / decompression device, an image compression method, an image decompression method, in which different quantization rates can be set for the attention area and other areas,
The present invention relates to a program that can be executed by a computer and a recording medium that records the program.

[0002]

2. Description of the Related Art Due to the progress of image input technology and its output technology, the demand for high definition for color still images has increased significantly in recent years. Taking a digital camera (DC) as an example of the image input device, the price of a high-performance electrically coupled device (CCD) having a pixel number of 3 million or more has been reduced, and the product is in a popular price range. Has also become widely used. Such high performance of the CCD largely depends on the progress of the silicon process or device technology, and has overcome the trade-off problem of miniaturization and reduction of S / N ratio. And it is said that this increasing number of pixels will continue for some time.

On the other hand, regarding image output / display devices, products in the hard copy field such as laser printers, ink jet printers and sublimation printers, and CRTs.
The high definition and low price of products in the soft copy field of flat panel displays such as LCDs, LCDs (liquid crystal display devices), and PDPs (plasma display devices) are remarkable.

Due to the market effect of such high-performance and low-priced image input / output products, the popularization of high-definition still images has started. It is expected that demand for high-definition still images will increase in all situations in the future. In fact, the development of technologies related to networks such as personal computers (PCs) and the Internet is accelerating these trends. Particularly in recent years, mobile devices such as mobile phones and notebook computers have become very popular at a very high speed, and the chances of transmitting or receiving images from any point using communication means are rapidly increasing. Against this background, it is inevitable that the demand for higher performance or more multifunctional image compression / decompression technology that facilitates the handling of high-definition still images will intensify in the future.

At present, as an image compression / decompression algorithm for facilitating the handling of such high-definition still images, JPEG (Joint Photographic) is used.
The Experts Group) is the most widely used. In addition, JPEG2000, which became certain to become an international standard in 2001, has not only a higher-performance algorithm than JPEG, but also a large number of functions and flexibility for various applications in parallel. As a result of achieving expandability, it is expected as a next-generation high-definition still image compression / decompression format succeeding JPEG.

FIG. 22 is a block diagram for explaining the basics of the JPEG algorithm. The JPEG algorithm includes a color space conversion / inverse conversion unit 50, a discrete cosine conversion /
The inverse transform unit 51, the quantization / inverse quantization unit 52, and the entropy encoding / decoding unit 53 are included. Usually, lossy encoding is used to obtain a high compression rate, and therefore compression / expansion of complete original image data, so-called lossless compression, is not performed in most cases. However, this lossy compression rarely causes a practical problem. Therefore, the JPEG method greatly contributes to suppressing the memory capacity required for compression / decompression processing or image data storage after compression, and shortening the time spent for data transmission / reception. Because of these advantages, JPEG has become the most popular still image compression / decompression algorithm at present.

FIG. 23 is a block diagram for explaining the basics of the JPEG2000 algorithm. JPEG2
000 algorithm is a color space conversion / inverse conversion unit 60,
It is composed of a two-dimensional wavelet transform / inverse transform unit 61, a quantization / inverse quantization unit 62, an entropy coding / decoding unit 63, and a tag processing unit 64.

As described above, the most widely used still image compression / decompression method is JPEG. However, the demand for high-definition still images has not stopped, and the technical limitations of the JPEG system are beginning to be seen. For example, block noises and mosquito noises, which were not so noticeable until now, have become noticeable as the original image has become higher in definition, and the image quality deterioration of the JPEG file has become a level that cannot be ignored. As a result, the improvement of image quality in a low bit rate, that is, a high compression rate region has come to be recognized as the most important issue in technological development. J
PEG2000 was born as an algorithm that can solve these problems. In the near future, it is expected that it will be used in combination with the currently mainstream JPEG format.

One of the biggest differences between the above-mentioned FIG. 22 and FIG. 23 is the conversion method. JPEG is a discrete cosine transform (DCT).
ine2000 is a discrete wavelet transform (DWT: Discrete).
Wavelet Transform) is used.
Compared with DCT, the advantage of DWT having good image quality in a high compression region is a major reason for its adoption. Another major difference is that in the latter, a functional block called a tag processing unit 64 is added to perform code formation at the final stage. Here, the codestream is generated and interpreted. The code stream has enabled JPEG2000 to realize various convenient functions. For example, FIG. 24 is a diagram showing an example of subbands at each decomposition level when the decomposition level is 3, and is an arbitrary hierarchy corresponding to the octave division hierarchy in the block-based DWT shown in FIG. Then, the compression / decompression process of the still image can be stopped.

The color space conversion / inverse conversion unit 50 and the color space conversion / inverse conversion unit 60 are often connected to the input / output portions of the original images in FIGS. 22 and 23. For example, the primary color R
RGB color system consisting of (red) / G (green) / B (blue) components and complementary color system Y (yellow) / M (magenta)
This corresponds to a portion that performs conversion from the YMC color system including each component of / C (cyan) to the YCrCb or YUV color system or the reverse conversion.

The JPEG2000 algorithm will be described below. Here, the definition of terms related to JPEG2000 is as follows: JPEG2000 PartI FDIS
(Final Draft International Standard). The definitions of typical terms are shown below. 1. code-block: A rectangular grouping of coefficie
nts from the same subband of a tile-component. decomposition level: A collection of wavelet su
bbands where each coefficient has the same spatial
impact or span with respect to the source compone
nt samples. These include the HL, LH, and HH subband
s of the same two dimensional subband decompositio
n.For the last decomposition level theLL subband i
s also included. precinct: A one rectangular region of a transfo
rmed tile-component, within each resolution level, u
sed for limiting the size of packets. layer: A collection of compressed image data fr
om coding passes of one, or more, code-blocks of a
tilecomponent. Layers have an order for encoding
and decoding that must be preserved. region of interest (ROI): A collection of coeffi
cients that are considered of particular relevance
by some user defined measure.

FIG. 25 is a diagram showing an example of each component of a tiled color image. The color image is
Generally, as shown in FIG. 25, each component 80, 81, 82 (here, RGB primary color system) of the original image is divided into rectangular areas (tiles) 80 _t , 81 _t , 82 _t . And the individual tiles, eg R00, R
01, ..., R15 / G00, G01, ..., G15 / B0
0, B01, ..., B15 are the basic units for executing the compression / decompression process. Therefore, the compression / decompression operation is performed independently for each component and each tile.

At the time of encoding, the data of each tile of each component is input to the color space conversion unit 60 shown in FIG. 23 and subjected to color space conversion, and then the two-dimensional wavelet conversion unit 61 performs the two-dimensional wavelet conversion. (Forward transform) is applied to spatially divide into frequency bands.

FIG. 24 described above shows subbands at each decomposition level when the decomposition level is 3. That is, the original image tile (0LL) (decomposition level 0 (70)) obtained by dividing the original image into tiles is subjected to the two-dimensional wavelet transform, and the decomposition level 1 (71)
Subbands shown in (1LL, 1HL, 1LH, 1HH)
To separate. Then, subsequently, the low frequency component 1LL in this hierarchy is subjected to a two-dimensional wavelet transform to separate the subbands (2LL, 2HL, 2LH, 2HH) shown in the decomposition level 2 (72). Similarly, the low-frequency component 2LL is similarly subjected to the two-dimensional wavelet transform, and the decomposition level 3 (7
3) Subbands (3LL, 3HL, 3LH, 3H)
H) is separated.

Further, in FIG. 24, the sub-bands to be coded at each decomposition level are shown in gray. For example, when the decomposition level is set to 3, the sub bands (3HL, 3LH, 3 shown in gray
HH, 2HL, 2LH, 2HH, 1HL, 1LH, 1H
H) is the target of encoding, and the 3LL subband is not encoded.

Next, the bits to be coded are determined in the designated coding order, and the quantizing unit 62 shown in FIG. 23 generates a context from the bits around the target bit. The wavelet coefficient after the quantization process is divided into non-overlapping rectangles called precincts for each subband. This was introduced to make the implementation use memory efficiently.

FIG. 26 is a diagram for explaining an example of the relationship between the precinct and the code block. The original image 90 is tiles 90 _t0 , 90 _t1 , 9 at decomposition level 1.
It is divided into four tiles of 0 _t2 and 90 _t3 . FIG. 26
As shown in, the precinct 90 _p4 is composed of three rectangular regions that are spatially coincident with each other.
_The same applies to _p6 . Here, the precinct numbers are assigned from 0 to 8 in raster order. Furthermore, each precinct is divided into non-overlapping rectangular code blocks. In this example, the code block is divided into 12 code blocks 0 to 11, and the code block 90 _b1 indicates the code block number 1, for example. This code block serves as a basic unit when performing entropy coding.

The entropy coding unit 63 shown in FIG. 23 described above performs coding on the tile of each component by probability estimation from the context and the target bit. In this way, the coding process is performed in tile units for all components of the original image. Finally, the tag processing unit 64 combines all the encoded data from the entropy coder unit into one code stream and adds a tag to the code stream. FIG. 27 is a diagram simply showing an example of the structure of the code stream. Tags called headers (main header 100 and tile part header 101, respectively) are provided at the beginning of the code stream and the beginning of the partial tiles forming each tile. Information is added, and thereafter, encoded data of each tile (bit stream 10
2) continues. Then, a tag is placed again at the end 103 of the code stream.

On the other hand, at the time of decoding, contrary to the time of encoding,
Image data is generated from the code stream of each tile of each component. A brief description will be given with reference to FIG. In this case, the tag processing unit 64 interprets the tag information added to the code stream input from the outside, decomposes the code stream into code streams of each tile of each component, and analyzes each code stream of each tile of each component. The decryption process is performed. The positions of the bits to be decoded are determined in the order based on the tag information in the code stream, and the dequantization unit 62
Then, a context is generated from a sequence of peripheral bits (which have already been decoded) at the target bit position. The entropy decoding unit 63 performs decoding by probability estimation from this context and code stream to generate a target bit, and writes it in the position of the target bit.

Since the data decoded in this way is spatially divided for each frequency band, a two-dimensional wavelet inverse transform unit 61 performs a two-dimensional wavelet inverse transform on the data, whereby each component of the image data is processed. Each tile is restored. The restored data is converted into the original colorimetric data by the color space inverse conversion unit 60.

Here, the coefficient values after the wavelet transform can be quantized and coded as they are,
In JPEG2000, in order to improve coding efficiency, coefficient values can be decomposed into bit plane units and the bit planes can be ranked for each pixel or code block.

FIG. 28 is a diagram simply showing an example of a procedure for ranking the bit planes. This example shows the original image 1
60 (32 × 32 pixels) is divided into four 16 × 16 pixel tiles (tile 160 _t0 , tile 160 _t1 , tile 160 _t2 ,
An illustration for a case divided by tiles 160 _t3), the size of the precinct and the code block of the decomposition level 1 is set to each 8 × 8 pixels and 4 × 4 pixels. The precinct and code block numbers are assigned in raster order. The pixel extension outside the tile boundary is performed by using the mirroring method, the wavelet transform is performed by the reversible (5,3) integer transform filter, and the wavelet coefficient value of decomposition level 1 is obtained.

Further, regarding the tile 160 _t0 (tile 0) / precinct 160 _p3 (precinct 3) / code block 160 _b3 (code block 3), a conceptual diagram showing an example of a typical layer structure is also shown. The converted code block 160 _w3 is the code block 160 _b3.
Is subjected to wavelet transform by a reversible (5,3) integer transform filter to obtain the wavelet coefficient value of decomposition level 1. Code block 1 after conversion
60 _w3 is a subband (1LL, 1HL, 1LH, 1H
H) and the wavelet coefficient value is assigned to each subband.

The layer structure is easy to understand when the wavelet coefficient values are viewed in the horizontal direction (bit plane direction).
One layer is composed of an arbitrary number of bit planes. In this example, layers 0, 1, 2, 3 are 1, respectively
It consists of 3, 1, and 3 bit planes. Then, a layer including a bit plane closer to the LSB is targeted for quantization earlier, and conversely, a layer closer to the MSB remains unquantized until the end. The method of discarding from the layer close to the LSB is called truncation, and the quantization rate can be finely controlled.

Further, in the case of the conventional JPEG compression / expansion format, the tile described in the above JPEG2000 may be read as a square block having a side of 8 pixels, which is subjected to two-dimensional discrete cosine transform.

Up to this point, a general still image has been described, but this technique can be extended to a moving image. That is, each frame of a moving image is composed of one still image, and video data is created (encoded) from these still images at a frame rate optimal for the application.
Or it can be displayed (decoded).
This is a function called the motion compression / decompression process of a still image. This system has a function not present in MPEG format video files that are widely used for moving images at the present time, that is, it has an advantage that high-quality still images can be handled on a frame-by-frame basis. It is starting to attract attention in the business field. Eventually, there is a high possibility that it will become popular with general consumers.

The specifications required for the compression / expansion algorithm for Motion still images differ greatly from the compression / expansion algorithm for general still images in processing speed.
This is because the frame rate that greatly affects the quality of moving images is determined. Therefore, in order to realize this function, at the present stage, it is limited to a method having high hardware dependency such as ASIC or DSP.
It will be necessary to wait for progress in process device technology in the semiconductor field and parallelizing compiler technology in the software field, etc., before sufficiently high-speed processing is possible with software.

However, according to the above prior art,
When the image is compressed and expanded, there is a problem that an important area, a focused area, and a non-important area are similarly compressed. For example, Japanese Patent Application Laid-Open No. 6-350989 (image data compression processing method) is an invention in view of the above problems and divides an image into a plurality of areas, assigns importance to each area, and assigns each area according to the importance. It controls the quantization rate of, but the importance of each region of the image is
The owners of the images vary widely and cannot reflect the needs of each individual. Further, the setting of the quantization rate for each area cannot be easily incorporated in the existing image compression apparatus.

Further, for example, Japanese Patent Laid-Open No. 2000-40142.
In the inventions disclosed in Japanese Patent Laid-Open No. 2001-285642, the data in the attention area is completely prioritized over the non-attention area by bit-shifting the bits in the attention area to the non-attention area. It was done. But,
In addition to the inability to finely control the quality of images in the attention area and the non-attention area, when the code size is reduced, a code may be created such that there is no data other than the attention area.

[0030]

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an important area of an image and an area of interest can be set by a user of the image, and the set area can be set. The image compression device, the image decompression device, the image compression / decompression device, the computer-executable program, and the program can be recorded. The present invention is made for the purpose of providing the recording medium.

Further, for the attention area and the non-attention area,
It is an object of the present invention to be able to finely set the resolution, to realize high compression while maintaining high image quality in accordance with the set resolution, and to shorten the processing time at that time.

Further, both the attention area and the non-attention area are reduced in data amount, and the attention area is highly accurate, the non-attention area is reduced in accuracy, and the accuracy is different for each area according to the user's request. By compressing the image, it is possible to set the attention area and the non-attention area for the image, set the attention degree for each set image area, and control the number of bits to be left according to the setting, The purpose is to perform image compression.

[0033]

According to a first aspect of the present invention, an image dividing means for dividing an image into a plurality of areas, an area specifying means for specifying one or a plurality of attention areas in the image, and the image dividing means. And a compression control means for controlling compression for each area based on the set quantization rate, and a compression rate setting means for setting the quantization rate for each area divided by It is characterized in that a quantization rate different from that of other areas can be set for the noted area.

According to a second aspect of the present invention, in the first aspect of the present invention, there is provided an attention degree setting means for setting an attention degree to the attention area designated by the area designation means, and the attention area having the attention degree set therein. On the other hand, a different quantization rate is set according to the degree of attention.

According to a third aspect of the invention, in the first or second aspect of the invention, at least two different quantization rates are set for the areas divided by the image dividing means. .

According to a fourth aspect of the present invention, in the third aspect of the invention, the quantization rate of the attention area designated by the area designating means is set higher than that of other areas. .

According to a fifth aspect of the invention, in the third aspect of the invention, the quantization rate of the attention area designated by the area designating means is set to be lower than that of other areas. .

According to a sixth aspect of the present invention, an image dividing means for dividing an image into a plurality of areas, an area specifying means for specifying one or a plurality of attention areas in the image, and each area divided by the image dividing means. It has bit number setting means for setting the number of bits and compression control means for controlling the compression for each area based on the set bit number, and the area of interest specified by the area specifying means is different from other areas. The feature is that different numbers of bits can be set.

According to a seventh aspect of the invention, in the sixth aspect of the invention, there is provided an attention degree setting means for setting the attention degree to the attention area designated by the area designation means, and the attention area having the attention degree set therein. On the other hand, a different number of bits is set according to the degree of attention.

The invention of claim 8 is characterized in that, in the invention of claim 6, at least two different bit numbers are set for the areas divided by the image dividing means according to the degree of attention. It is a thing.

According to a ninth aspect of the present invention, in the eighth aspect, the number of bits of the attention area designated by the area designating unit is set to be larger with respect to other areas as the degree of attention is higher. It is a feature.

According to a tenth aspect of the present invention, in the invention of the eighth aspect, the number of bits of the attention area designated by the area designating unit is set to be smaller with respect to other areas as the attention degree is higher. It is a feature.

The invention of claim 11 is based on claims 1 to 10.
In any one of the inventions, the area designating means designates one or a plurality of target areas from the areas divided by the image dividing means.

The invention of claim 12 relates to claims 1 to 11.
In any one of the inventions, the area designating means designates one or a plurality of areas of interest in accordance with a portion of interest in the image.

According to a thirteenth aspect of the present invention, an image dividing means for dividing an image into a plurality of areas, a resolution setting means for setting a resolution for each area divided by the image dividing means, and an area for each area according to the set resolution. It has a compression control means for controlling compression at different resolutions.

According to a fourteenth aspect of the present invention, in the thirteenth aspect, at least two different resolutions are set for the areas divided by the image dividing means.

According to a fifteenth aspect of the invention, in the fourteenth aspect of the invention, the resolution of at least one of the areas divided by the image dividing means is set higher than that of the other areas. It is a thing.

According to a sixteenth aspect of the present invention, in the fourteenth aspect, the resolution of at least one area of the area divided by the image dividing means is set lower than that of the other areas. It is a thing.

The invention of claim 17 is based on claims 13 to 1.
In the invention of any one of 6), the calculation of the wavelet coefficient is omitted according to the resolution set by the resolution setting means.

The eighteenth aspect of the invention is characterized in that, in the seventeenth aspect of the invention, the wavelet coefficient for which the calculation is omitted is a high frequency component.

The invention of claim 19 is characterized in that, in the invention of claim 18, an average calculation is used for the calculation of the low frequency component of the wavelet coefficient of the specific region.

The twentieth aspect of the invention is characterized in that, in the eighteenth aspect of the invention, a specific element is extracted and used for the calculation of the low frequency component of the wavelet coefficient of the specific region.

The invention of claim 21 is the invention of claims 1 to 20.
In any one of the inventions described above, when the image is divided into a plurality of regions, the tiles are used for division.

The invention of claim 22 is based on claims 1 to 20.
In any one of the inventions described above, when the image is divided into a plurality of regions, the image is divided using a precinct.

The invention of claim 23 is based on claims 1 to 20.
In any one of the inventions, when the image is divided into a plurality of regions, the division is performed using a code block.

The invention of claim 24 is based on claims 1 to 20.
In any one of the inventions, when the image is divided into a plurality of regions, the image is divided in pixel units.

The invention of claim 25 is based on claims 1 to 20.
In any one of the inventions, when dividing an image into a plurality of regions, any two or more of tile, precinct, code block, and pixel unit are combined and divided.

The invention of claim 26 is based on claims 1 to 25.
In any one of the inventions 1 to 3, when the image is divided into a plurality of regions, the region is set in accordance with a target portion in the image,
The feature is that the image is divided according to the set area.

The invention of claim 27 is characterized in that, in the invention of claim 26, the boundary of the divided areas is a tile boundary.

According to a twenty-eighth aspect of the present invention, in the twenty-sixth aspect, the boundary of the divided areas is a precinct boundary.

The invention of claim 29 is characterized in that, in the invention of claim 26, the boundary of the divided areas is a code block boundary.

The invention of claim 30 is characterized in that, in the invention of claim 26, the boundaries of the divided areas are any two or more of a tile boundary, a precinct boundary and a code block boundary. Is.

According to a thirty-first aspect of the present invention, an image dividing means for dividing an image into a plurality of areas, a bit plane decomposition means for dividing the area divided by the image dividing means into bit planes, and the decomposed bit planes And a compression control unit that controls compression for each area based on the attention level set for the bit plane, and the attention degree setting unit sets the attention level according to the attention level set for the bit plane. The feature is that a compressed image is generated.

According to a thirty-second aspect of the present invention, an image dividing unit that divides an image into a plurality of regions, a bit plane decomposition unit that decomposes the region divided by the image dividing unit into bit planes, and the decomposed bit planes Degree setting means for setting the degree of attention, compression control means for controlling compression for each region based on the degree of attention set in the bit plane, and tag processing control means for controlling tag processing based on the degree of attention And a compressed image is generated according to the degree of attention set in the bit plane.

The invention of claim 33 relates to claim 31 or 3
In the second invention, the attention level is set for the bit planes corresponding to the respective areas divided by the image dividing means.

According to a thirty-fourth aspect of the present invention, in the thirty-third aspect, at least two different attention levels are set for the bit planes corresponding to the areas divided by the image dividing means. It is a thing.

According to a thirty-fifth aspect of the present invention, in the thirty-third aspect, the bit plane of at least one image area divided by the image dividing means has a higher degree of attention than a degree of attention set in another image area. It is characterized by setting many.

According to a thirty-sixth aspect of the present invention, in the thirty-third aspect, the bit plane of at least one image area divided by the image dividing means has a higher attention degree than the attention degree set in another image area. It is characterized by setting a small number.

The invention of claim 37 is based on claims 1 to 36.
In any one of the inventions described above, the two-dimensional discrete wavelet transformer, the quantizer, and the entropy encoder are combined.

The invention of claim 38 is based on claims 31 to 3.
An image decompression device for decompressing a compressed image compressed by using the image compression device according to any one of 7, wherein the compressed image is decompressed based on the attention-related information included in the compressed image. It is characterized by.

The invention of claim 39 relates to claims 13 to 2.
An image decompression device for decompressing a compressed image compressed using the image compression device according to any one of 5, wherein the compressed image is decompressed based on the information regarding the resolution included in the compressed image. It is a feature.

The invention of claim 40 is based on claims 1 to 37.
An image decompression device for decompressing a compressed image compressed using the image compression device according to any one of 1,
It is characterized in that the compressed image is decompressed based on the information of the area divided by the image dividing means.

According to a forty-first aspect of the invention, the image is divided into a plurality of areas, a quantization rate is set for each of the divided areas, and compression is performed by controlling the compression for each area based on the set quantization rate. An image decompression device for decompressing a compressed image that has been compressed is characterized by decompressing a compressed image based on information of divided areas included in the compressed image.

According to a forty-second aspect of the invention, the image is divided into a plurality of areas, the number of bits is set for each of the divided areas, and compression is controlled for each area based on the set number of bits for compression. An image decompression device for decompressing a compressed image, which is characterized in that the compressed image is decompressed based on information of a divided area included in the compressed image.

The invention of claim 43 is based on claims 38 to 4.
The invention of any one of 2) is characterized in that it is configured by a combination of a two-dimensional discrete wavelet inverse transformer, an inverse quantizer, and an entropy decoder.

According to a forty-fourth aspect of the invention, the image is divided into a plurality of areas, a quantization rate is set for each of the divided areas, and compression is performed by controlling compression for each area based on the set quantization rate. An image compression / decompression device for decompressing the compressed compressed image, comprising a two-dimensional discrete wavelet transform / inverse transformer,
It is characterized by being configured by a combination of a quantizer / inverse quantizer and an entropy encoder / decoder.

According to a forty-fifth aspect of the present invention, the image is divided into a plurality of areas, the number of bits is set for each of the divided areas, and compression is performed by controlling the compression for each area based on the set number of bits. An image compression / decompression device for decompressing the compressed compressed image, comprising a two-dimensional discrete wavelet transform / inverse transformer,
It is characterized by being configured by a combination of a quantizer / inverse quantizer and an entropy encoder / decoder.

According to the forty-sixth aspect of the present invention, the image is divided into a plurality of areas, the divided areas are decomposed into bit planes, the degree of attention is set to the decomposed bit planes, and the degree of attention is set based on the set degree of attention. The compression is controlled for each area, and the tag information is controlled based on the set degree of attention to generate a compressed image,
An image compression / decompression device for decompressing the generated compressed image, comprising a combination of a two-dimensional discrete wavelet transform / inverse transformer, a quantizer / inverse quantizer, and an entropy encoder / decoder. It is characterized by being present.

According to a forty-seventh aspect of the present invention, means for dividing an image into a plurality of areas, means for setting a resolution for each of the divided areas, and means for controlling compression at a different resolution for each area according to the set resolution. And a means for expanding the compressed compressed image based on the resolution set in the compressed image.

The invention of claim 48 is based on claims 1 to 37.
The image compression device according to any one of 1 to 3, further comprising a combination of a two-dimensional discrete wavelet transform / inverse transformer, a quantizer / inverse quantizer, and an entropy encoder / decoder. It is a feature.

According to a 49th aspect of the invention, an image dividing step of dividing an image into a plurality of areas, an area specifying step of specifying one or a plurality of attention areas in the image, and a quantization rate for each of the divided areas are set. Quantization rate setting step to set,
A compression control step of controlling compression for each area based on the set quantization rate, and a quantization rate different from that of other areas can be set for the attention area. .

According to a fiftieth aspect of the present invention, an image dividing step of dividing an image into a plurality of areas, an area specifying step of specifying one or a plurality of attention areas in the image, and a bit number for each of the divided areas are set. Bit number setting step,
A compression control step of controlling compression for each area based on the set number of bits, and a bit number different from those of other areas can be set for the attention area.

According to a fifty-first aspect of the invention, an image division step of dividing an image into a plurality of areas, a bit plane decomposition step of decomposing the divided areas into bit planes, and a degree of attention to the decomposed bit planes An attention level setting step for setting, a compression control step for controlling compression for each area based on the attention degree set for the bit plane, and a tag processing control step for controlling tag processing based on the attention degree. The compressed image is generated according to the attention level set in the bit plane.

According to a fifty-second aspect of the invention, an image dividing step of dividing an image into a plurality of areas, a resolution setting step of setting a resolution for each of the divided areas, and a compression with a different resolution for each area according to the set resolution. And a compression control step for controlling.

According to a fifty-third aspect of the invention, the image is divided into a plurality of areas, a quantization rate is set for each of the divided areas, and compression is performed by controlling compression for each area based on the set quantization rate. An image decompression method for decompressing a compressed image that has been compressed is characterized in that the compressed image is decompressed based on information of divided areas included in the compressed image.

According to the invention of claim 54, the image is divided into a plurality of areas, the number of bits is set for each of the divided areas, and the compression is controlled for each area based on the set number of bits for compression. An image decompression method for decompressing a compressed image is characterized in that the compressed image is decompressed based on information of a divided area included in the compressed image.

According to the 55th aspect of the present invention, the image is divided into a plurality of areas, the attention level is set for each bit plane of the divided area, and the compression is controlled for each area based on the set attention level. An image decompression method for decompressing a compressed image generated by controlling tag processing based on the set attention level,
It is characterized in that a compressed image is expanded based on the information regarding the degree of attention included in the compressed image.

The invention of claim 56 is based on claims 1 to 37.
A computer functioning as the image compression apparatus according to any one of claims 1 to 3 or each unit of the image compression apparatus, or as the image decompression apparatus according to any one of claims 38 to 43 or each unit of the image decompression apparatus. It is a program to let you.

The invention of claim 57 is based on claims 1 to 37.
A computer functioning as the image compression apparatus according to any one of claims 1 to 3 or each unit of the image compression apparatus, or as the image decompression apparatus according to any one of claims 38 to 43 or each unit of the image decompression apparatus. It is a computer-readable recording medium in which a program for causing the program is recorded.

[0090]

1 is a block diagram showing an example of the configuration of an image compression / decompression device for still images according to an embodiment of the present invention, in which 10 is a color space conversion / inverse conversion unit. , 11 are first components, and the first component 11 includes a wavelet transform / inverse transform unit 11a, a quantization rate selection unit 11b, a quantization / inverse quantization unit 11c, and an entropy coding / decoding unit 11d. The second component 12 is a wavelet transform / inverse transform unit 12
a, a quantization rate selection unit 12b, a quantization / inverse quantization unit 12
c, an entropy coding / decoding unit 12d, and a third
The component 13 includes a wavelet transform / inverse transform unit 13a, a quantization rate selection unit 13b, a quantization / inverse quantization unit 1
3c, comprising an entropy coding / decoding section 13d, 1
Reference numeral 4 is a code stream processing unit, and 15 is a quantization rate setting unit.

In the image compression / decompression device shown in FIG. 1, one conversion / inverse conversion unit in each processing block is shown so as to be able to deal with image compression and expansion. You may take the form divided into parts.
By doing so, it becomes possible to independently use the image compression device and the image expansion device. Further, the image compression / decompression device of the present invention or the compressed image decompressed by the image decompression device is not limited to the compressed image compressed by the image compression device of the present invention, and the image is divided into a plurality of regions and divided. It is possible to set the quantization rate for each area, control the compression for each area based on the set quantization rate, and apply the compression image to all compressed images.

According to the present invention, the method using the DWT in JPEG2000 is introduced into a concrete apparatus configuration for realizing the image compression / expansion processing, and therefore, the present invention can be continuously used now and in the future. You can

The image compression / decompression device shown above is intended for still images. Processing is component 11,
It progresses in parallel every 12 and 13. The color space here is
RGB or YUV. The case of compression will be described below. First, in the color space conversion / inverse conversion unit 10, the color space is converted into the RGB or YUV color space, and the processing is performed in parallel for each of the components 11, 12, and 13. Here, the first component 11 will be described as an example, but other second components 12 and third components 13 are also subjected to similar processing. First, the image subjected to color space conversion is divided into tile areas, and wavelet conversion processing is performed by the wavelet conversion / inverse conversion unit 11a for each divided tile area. Here, the divided area is
Region division is performed using tiles, and precincts and code blocks are segmented after wavelet transform processing. Also, these tiles, precincts,
It is also possible to combine code blocks to divide into regions.

Next, the quantization rate selecting unit 11b selects the quantization rate for each wavelet-transformed area. When the user specifies the attention area and sets the quantization rate for each area, The region of interest in the image can be designated by the quantization rate setting unit 15, and a quantization rate different from that of other regions can be set for the designated region of interest. In this way, after the quantization rate is determined for each area of the image, the wavelet coefficient is quantized by the quantization / inverse quantization unit 11c, and further encoded by the entropy coding / decoding unit 11d. Finally, the codestream processing unit 14 generates a codestream.

According to the present invention, the quantization rate is set according to the area selected by the user, and the compression is controlled based on the quantization rate. Therefore, it is possible to realize the image compression that suits the individual taste. .

FIG. 2 is a block diagram for explaining the details of a configuration example of an image compression / expansion apparatus to which the present invention is applied. In the figure, 20 is a compression / expansion apparatus, and the compression / expansion apparatus 20 is a color space. The transform / inverse transform unit 20a, the two-dimensional wavelet transform / inverse transform unit 20b, the quantization / inverse quantization unit 20c, the entropy coding / decoding unit 20d, and the tag processing unit 20e, 21 is an image dividing unit, and 22 is a compression unit. Control means, 23 is a quantization rate setting means, and 24 is a region designating means. The compression / decompression device 20 performs the same processing as the processing in each component shown in FIG. 1 described above. Here, the tag processing unit 2
0e is added as one of the features of JPEG2000, and is a functional block for generating and interpreting a codestream.

The original image is color space converted by the color space conversion / inverse conversion unit 20a, and then divided into a plurality of regions by the image dividing means 21. When the image is divided, one of the tile, the precinct, and the code block is used for division, but any two or all of these may be combined and divided. It is also possible to specify a target portion in the image, set a target area in accordance with the target portion, and divide the image by the set area. In this case, the boundary of the divided area is a tile boundary, a precinct boundary, or a code block boundary, or any two or three of them, depending on the division unit of the area.
All three are the boundaries of the area.

According to the present invention, JP
Tile, precinct, which is the standard specification of EG2000,
The compression control can be easily realized by using the code block.

The area specifying means 24 specifies the area of interest in the image. Specifically, for example, a target portion in an image displayed on a display or the like can be designated as a target region by using an input means such as a mouse or a keyboard, and can be designated as a target region in accordance with the designated range. At this time, the attention area to be designated may be one or more. As another method, the attention area may be designated from the areas divided by the image dividing unit 21 using tiles, precincts, or code blocks. In this case, for example, for the image displayed on the display or the like, a number is assigned to each of the divided areas, and the number corresponding to the attention area is designated by the mouse or the keyboard, whereby the attention area can be designated. At this time, the attention area specified is 1
One or more may be used.

According to the present invention, when an image is divided into a plurality of areas, the image is divided according to the area set by the user, so that efficient image compression control is possible. Further, by using tiles, precincts, and code blocks that are the standard specifications of JPEG2000, compression control can be easily realized.

The quantization rate setting means 23 sets the quantization rate for each area. Specifically, the attention level is set in the attention area designated by the area designating unit 24. Here, the degree of attention is a standard for controlling compression of the attention area in the image, for example, reducing the quantization rate in a region where higher definition is desired, and the attention area is set by using the attention degree setting means described later. It can be set appropriately for each user's preference. By setting different quantization rates according to the degree of attention set by the user, it is possible to perform compression while maintaining image quality accuracy. In addition, as a form that is often used, two kinds of quantization rates can be set for divided areas. In this case, the quantization rate may be set such that the attention area and the other areas have different quantization rates. That is, if the quantization rate of the attention area is set higher than that of the other area, the attention area becomes coarser in terms of image quality, and if the quantization rate of the attention area is set lower than that of the other area, Is higher in terms of image quality. In this way, the quantization rate can be appropriately set according to the preference of the user.

According to the present invention, by setting a plurality of quantization rates according to the degree of attention set in the attention area, it is possible to perform compression control while maintaining high image quality. Further, since at least two kinds of quantization rates are set for the divided areas of the image, it is possible to easily realize the compression control that reflects the user's preference with a simple configuration.

The compression control means 22 controls the compression for each area based on the set quantization rate. The compression-controlled area is quantized by the quantization / inverse quantization section 20c based on the quantization rate, entropy coded by the entropy coding / decoding section 20d, and code stream processed by the tag processing section 20e. .

FIG. 3 is a block diagram showing a configuration example of an image compression / decompression device for a moving image according to the present invention. In the figure, 30 is a moving image compression / decompression device for a moving image.
The moving picture compression / decompression device 30 includes frame control means 30.
a, a still image compression / decompression unit 30b, and a compressed moving image processing unit 30c. When the moving image compression / decompression device 30 receives the moving image, the frame control unit 30a controls the frame, and the still image compression / decompression unit 30b is performed for each frame.
Pass to. The still image compression / decompression means 30b compresses / decompresses a still image as shown in FIG. The compressed image is created by the compressed moving image processing means 30c. At this time, a plurality of still image compression / decompression means 30b may be provided to improve the processing speed. In addition, the quantization rate for each area can be set for all images, and it can be set when changes occur, or the movement of the target object can be automatically calculated from the interframe difference data, etc. It is also possible to use a method of detecting the above, and setting the quantization rate.

FIG. 4 is a flow chart for explaining an example of image compression processing to which the present invention is applied. In this example, when the image is divided into a plurality of areas, a case of using tiles is shown as a representative example, but the same processing can be performed even when a precinct or a code block is used. First, when the image compression / decompression device of the present invention receives an input image, the image compression / decompression device according to the input image color space
If it is determined that the color space conversion to B, YUV, or the like is to be performed (step S1), and it is determined to be necessary (YE
In the case of S), color conversion is performed (step S2). In step S1, it is naturally conceivable that the input image has already been subjected to color space conversion, or if compression is performed without performing color space conversion. In this case (NO), the process proceeds to step S3. Next, it is determined whether or not the image is divided into areas (step S3). Here, the area division determination may be performed automatically according to the size of the image, or as a command given by the user. When it is determined that image segmentation is necessary (YES)
Divides the image (step S4). In this example, the image is divided using tiles. If it is determined that image division is not necessary (NO), step S
Go to 5.

Next, it is judged whether or not the attention area in the image is designated (step S5). When the attention area is designated (in the case of YES), the attention area in the image is designated (step S6). Specifically, for example, a target portion in an image displayed on a display or the like can be designated as a target region by using an input means such as a mouse or a keyboard, and can be designated as a target region in accordance with the designated range. At this time, the attention area to be designated may be one or more. Further, as another method, the attention area may be designated from the areas divided by using tiles, precincts, or code blocks. In this case, for example, for the image displayed on the display or the like, a number is assigned to each of the divided areas, and the number corresponding to the attention area is designated by the mouse or the keyboard, whereby the attention area can be designated. When the attention area is not specified (NO),
Control goes to step S7.

Next, in step S7, it is determined whether or not the processing has been completed for all areas (step S7).
7) If not completed (in the case of NO), wavelet transform is performed for each divided area (step S8). At this time, the wavelet transform may be performed after dividing the image into tiles in step S4 described above. In addition, for precincts and code blocks, area division is performed after wavelet transform processing.
Next, it is determined whether the quantization rate needs to be changed (step S9), it is determined that the quantization rate needs to be changed, and the user has designated the quantization rate for each area (YE
In the case of S), the quantization rate is set for the corresponding divided area (step S10). When the quantization rate is not designated (NO), the reference quantization rate is set for each area (step S11).

Next, quantization is performed according to the set quantization rate (step S12), and encoding is performed (step S1).
3). When the encoding is completed, the counter for counting the divided areas is incremented (step S14), and the process returns to step S7. Hereinafter, the processing from step S7 to step S14 is repeated until the processing is completed for all areas.
When the processing is completed for all areas (YES in step S7), a code stream is generated (step S15),
The process ends.

FIG. 5 is a diagram showing an embodiment in which the present invention is applied to a still gray scale image. In the case where the region including the human face shown in FIG. 5A is set as a region of interest and the quantization rate is set low, a unit when dividing the region will be described as a tile. As shown in FIG. 5A, the original image 40
Is divided into a plurality of tiles and, for example, when tiles are numbered from 0 to 15 in raster order as shown in FIG. 25, the tile number 5 corresponding to the black-painted portion 41 shown in FIG. , 6, 9 and 10 are kept low, the accuracy of the region including the human face is maintained and the original image 4
High compression of 0 can be realized.

FIG. 6 is a diagram showing another embodiment in which the present invention is applied to a still gray scale image. 6A shows division of an image area in the case where the user designates the human face portion shown in FIG. 6A as the attention area and wants to make the quantization rate of the attention area different from other areas. . In this example, tiles are used as a unit for area division. By setting the tile size so as to cover the human face portion of the original image 40, efficient compression control is possible. In the case of this example, the original image 40 can be highly compressed by reducing the quantization rate only for the tile number 10 corresponding to the black-painted portion 42 shown in FIG. 6B. Of course, it is not necessary to cover the face portion with one tile, and it may be covered with a plurality of tiles.

FIG. 7 is a diagram showing an example in which a precinct is applied to area division. Similar to the example of the tile shown in FIG. 5, a case where the quantization rate is set low with the region including the human face shown in FIG. 7A as the region of interest will be described as a precinct as a unit at the time of region division. . As shown in FIG. 7A, the original image 40 is divided into a plurality of precincts, and the quantization rate of the precinct corresponding to the black-painted portion 43 shown in FIG. It is possible to realize high compression of the original image 40 while maintaining the accuracy of the included area.

FIG. 8 is a diagram showing an example in which a code block is applied to area division. Similar to the example of the tile shown in FIG. 5, the case where the region including the human face shown in FIG. 8A is set as the attention region and the quantization rate is set low, the unit at the time of region division is described as a code block. To do. FIG. 8 (A)
As shown in FIG. 7, the original image 40 is divided into a plurality of code blocks, and the quantization rate of the code block corresponding to the black-painted portion 44 shown in FIG. It is possible to realize high compression of the original image 40 while maintaining accuracy.

In the present embodiment, how to reduce the quantization rate, that is, the quantization step size is not specified, but it may be set to a certain value or the quantization step size may be set to 1, that is, lossless. It is possible.

FIG. 9 is a flow chart for explaining the image decompression method according to the embodiment of the present invention. First, tag information is interpreted based on the tile division information included in the compressed image (step S21), entropy decoding (step S22), inverse quantization (step S23),
Inverse wavelet transformation (step S24) is performed to restore the tiles (step S25), the restored tiles are integrated and inverse color space transformation is performed (step S26), and a decompressed image is generated. Here, a compressed image that is expanded by the image expansion method shown in this example is divided into a plurality of areas, a quantization rate is set for each divided area, and compression is performed for each area based on the set quantization rate. Compressed images that are compressed by controlling the image, or divide the image into multiple areas, set a layer for each divided area, and control the compression for each area based on the bit plane division for the set layer. Compressed general image or divided the image into multiple areas, decompose the divided area into bit planes, set the degree of attention to the decomposed bit planes, and control the compression for each area based on the set degree of attention The target is all compressed images that have been compressed. Furthermore, this image decompression method divides the image into a plurality of areas, sets the resolution for each divided area,
It is also possible to control the compression for each area according to the set resolution and target the entire compressed image compressed.

According to the present invention, it is possible to decompress a compressed image compressed by the image compression device. Further, the image is divided into a plurality of regions, a quantization rate is set for each of the divided regions, compression is controlled for each region based on the set quantization rate, and the compressed image that has been compressed is expanded. Is possible. Further, the image is divided into a plurality of areas, bit plane division is set for each divided area, compression is controlled for each area based on the set bit plane division, and the compressed image is expanded. Is possible. It is also possible to divide an image into a plurality of areas, set a degree of attention to the bit plane for each divided area, and control compression based on the set degree of attention to decompress a compressed image that has been compressed. It will be possible. It is also possible to divide an image into a plurality of areas, set a resolution for each of the divided areas, control compression for each area according to the set resolution, and decompress the compressed image that has been compressed.

FIG. 10 is a block diagram showing another configuration example of an image compression / decompression device for a still image according to an embodiment of the present invention. In the figure, 110 is a color space conversion / inverse conversion unit, 111 Is a first component, and the first component 111 is a wavelet transform / inverse transform unit 111.
a, a quantization / inverse quantization unit 111b, a bit plane division unit 111c, and an entropy encoding / decoding unit 111d. Similarly, the second component 112 includes a wavelet transform / inverse transform unit 112a and a quantization / inverse quantization unit. 11
2b, a bit plane division unit 112c, and an entropy coding / decoding unit 112d, and a third component 1
13 is a wavelet transform / inverse transform unit 113a, a quantization / inverse quantization unit 113b, and a bit plane division unit 113.
c, consisting of entropy coding / decoding section 113d, 1
Reference numeral 14 is a code stream processing unit, and 115 is a bit plane setting unit.

In the image compression / expansion device shown in FIG. 10, the conversion / inverse conversion unit in each processing block is shown as one so as to be capable of compressing and expanding the image. You may take the form divided into parts. By doing so, it becomes possible to independently use the image compression device and the image expansion device. Further, the image compression / decompression device of the present invention, or the compressed image decompressed by the image decompression device is not limited to the compressed image compressed by the image compression device of the present invention, and the image is divided into a plurality of regions,
A bit plane can be divided and set for each divided area, and compression can be controlled for each area based on the divided and set bit planes to apply to compressed images in general.

The image compression / decompression device shown above is for a still image. Processing is component 11
Every 1,112,113 is advanced in parallel. The color space here is RGB or YUV.

The case of compression will be described below. First, in the color space conversion / inverse conversion unit 110, the color space is converted into RGB or YUV color space, and the component 11
Processing is advanced in parallel for each of 1,112,113. Here, the first component 111 will be described as an example, but other second components 112 and third components 11 will be described.
The same processing is performed for 3. The color space converted image is
It is divided into tiles, and wavelet transform processing is performed for each tile by the wavelet transform / inverse transform unit 111a. Next, the quantization / inverse quantization unit 111b quantizes the wavelet-transformed coefficients.

The bit-plane dividing unit 111c divides the bit-plane consisting of the quantized wavelet coefficients into regions. When the user designates a region of interest and the region is divided into bit-planes, the bit-planes are divided into bit-planes. It is possible to specify the attention area in the image by the setting unit 115, and set the division of the bit plane different from other areas for the designated attention area. Thus
After setting the bit plane division for each area of the image, the wavelet coefficients are further entropy code /
Encoding is performed by the decoding unit 111d. Here, the area to be divided is divided into areas using any of tiles, precincts, and code blocks. Further, these tiles, precincts, and code blocks can be combined to divide into regions. Finally, the codestream processing unit 114 generates a codestream.

According to the present invention, the bit plane division is set according to the area selected by the user, and the compression is controlled based on the bit plane division. Therefore, it is possible to realize the image compression that suits the individual taste. . Also, by setting a plurality of bit plane divisions according to the degree of attention set in the attention area, it becomes possible to perform compression control while maintaining high image quality.

FIG. 11 is a block diagram for explaining the details of another structural example of the image compression / expansion apparatus to which the present invention is applied. In the figure, reference numeral 120 denotes the compression / expansion apparatus.
Reference numeral 20 denotes a color space conversion / inverse conversion unit 120a, a two-dimensional wavelet conversion / inverse conversion unit 120b, a quantization / inverse quantization unit 120c, a bit plane division unit 120d, an entropy encoding / decoding unit 120e, and a tag processing unit 120f. , 121 is an image division means, 122 is a bit plane division control means, 123 is a bit plane division designation means, 1
Reference numeral 24 is an area designating means. The compression / expansion device 120 performs the same processing as the processing in each component shown in FIG. 10 described above. Here, the tag processing unit 120f uses JP
It is added as one of the features of EG2000, and is a functional block for generating and interpreting a codestream.

The original image is a color space conversion / inverse conversion unit 120a.
After the color space conversion, the two-dimensional wavelet conversion is performed, the quantization / inverse quantization unit 120c performs quantization based on the quantization rate, and the image dividing unit 121 divides the image into a plurality of regions. When the image is divided, one of the tile, the precinct, and the code block is used for division, but any two or all of these may be combined and divided. It is also possible to specify a target portion in the image, set a target area in accordance with the target portion, and divide the image by the set area. In this case, depending on the division unit of the area, the boundary of the divided area is any one of a tile boundary, a precinct boundary, and a code block boundary, or any two or all of them are It becomes the boundary of the area.

A region of interest in the image is designated by the region designating means 124. Specifically, for example, a target portion in an image displayed on a display or the like can be designated as a target region by using an input means such as a mouse or a keyboard, and can be designated as a target region in accordance with the designated range. At this time, the attention area to be designated may be one or more. As another method, the attention area may be designated from the areas divided by the image dividing unit 121 using tiles, precincts, or code blocks. In this case, for example, for the image displayed on the display or the like, a number is assigned to each of the divided areas, and the number corresponding to the attention area is designated by the mouse or the keyboard, whereby the attention area can be designated. Also in this case, one or more attention areas may be designated.

The quantized wavelet coefficient is divided into layers, which are a collection of bit planes, in the bit plane division unit 120d. This division is set by the bit plane division control unit 122 in accordance with the division of the bit plane for each area designated by the bit plane division designation unit 123. Specifically, the degree of attention is set in the attention area designated by the area designating unit 124. Here, the degree of attention is a guideline for controlling compression of the attention area in the image, for example, by placing more bits in a layer that is prioritized in an area where higher definition is desired, and the preference of the user for each attention area. It can be appropriately set according to. By setting different bit plane divisions according to the degree of attention set by the user, it is possible to perform compression while maintaining image quality accuracy.

In addition, as a form that is often used,
Two types of bit plane divisions can be set for the divided areas. In this case, the bit plane divisions of the attention area and other areas may be set to be different. That is, the bit plane of the attention area (the number of bits)
If more areas are included in the priority layer, the attention area has higher definition in terms of image quality, and if the attention area has a smaller number of bit planes (bits), the attention area is more However, the image quality is rough. In this way, the bit plane division can be set appropriately according to the preference of the user.

The compression control means 22 controls the compression by controlling the number of bits included in the layer for each area based on the set bit plane division. The compression-controlled area is entropy-encoded by the entropy coding / decoding section 120e, and code-stream processed by the tag processing section 120f.

According to the present invention, since at least two kinds of bit plane division are performed on the divided area of the image,
With a simple configuration, compression control that reflects the user's taste can be easily realized. Also, for image area division, JPE
By using tiles, precincts, and code blocks that are standard specifications of G2000, it is possible to easily realize compression control. Further, when the image is divided into a plurality of areas, the image is divided according to the area set by the user, so that efficient image compression control is possible.

The image compression / decompression device shown in FIG. 11 can be applied to a moving image by adopting the same configuration as the moving image compression / decompression device 30 shown in FIG. At this time, it is possible to set the bit plane division for each area for all images, or to set the method when a change occurs, or to automatically move the target object from the inter-frame difference data etc. It is also possible to use a method of detecting the same and setting the bit plane division.

FIG. 12 is a flow chart for explaining an example of another image compression process to which the present invention is applied. In this example, a case where a precinct is used to divide an image into a plurality of areas is shown as a representative example, but the same processing can be performed when tiles or code blocks are used. First, when the image compression / decompression device of the present invention receives an input image, it divides the image into tiles (step S31), and determines whether to perform color space conversion into RGB, YUV, or the like according to the color space of the input image. If it is determined (step S32) and it is determined to be necessary (YES), color conversion is performed (step S33). In step S32, it is naturally conceivable that the input image has already undergone color space conversion, or if compression is performed without performing color space conversion. In this case (NO), the process proceeds to step S34. Next, a wavelet transform is applied (step S34) and quantization is performed (step S3).
5). Further, it is determined whether or not the image is divided into areas (step S36). Here, the area division determination may be performed automatically according to the size of the image, or as a command given by the user. If it is determined that image division is necessary (YES), image division is performed (step S37). Image division is performed using a precinct in this example. If it is determined that the image division is not necessary (NO), or if the division unit is a tile, the process proceeds to step S38.

Next, it is judged whether or not the attention area in the image is designated (step S38). If the attention area is designated (in the case of YES), the attention area in the image is designated (step S39). Specifically, for example, a target portion in an image displayed on a display or the like can be designated as a target region by using an input means such as a mouse or a keyboard, and can be designated as a target region in accordance with the designated range. At this time, the attention area to be designated may be one or more.
Alternatively, you can use tiles or precincts,
Alternatively, the attention area may be designated from the areas divided using the code block. In this case, for example, for the image displayed on the display or the like, a number is assigned to each of the divided areas, and the number corresponding to the attention area is designated by the mouse or the keyboard, whereby the attention area can be designated. When the attention area is not designated (in the case of NO), the process proceeds to step S40.

Next, in step S40, it is judged whether or not the processing has been completed for all the areas (step S40), and if not completed (in the case of NO),
For each of the divided areas, it is determined whether the bit plane division is set (step S41). When it is determined that the bit plane division needs to be changed, and the user has designated the bit plane division for each area (in the case of YES), the bit plane division is set for the corresponding division area. (Step S42). If the bit plane division is not designated (NO), the reference bit plane division is set for each area (step S43).

Next, encoding is performed (step S44).
When the encoding is completed, the counter for counting the divided areas is incremented (step S45), and the step S
Return to 40. Hereinafter, the processing from step S40 to step S45 is repeated until the processing is completed for all areas, and when the processing is completed for all areas (in the case of YES in step S40), code stream processing is performed (step S4).
6), the process ends.

Here, in FIG. 5 described above, the present invention is applied to a still gray scale image, and many bit planes are set as low layers with the region including the human face shown in FIG. 5A as the region of interest. In the case of arranging the tiles, the unit for dividing the area will be described as a tile. When the original image 40 is divided into a plurality of tiles as shown in FIG. 5A and tiles are numbered from 0 to 15 in raster order as shown in FIG. 22, for example, when the tiles are numbered as shown in FIG. Black-painted part 4 shown
The tiles with tile numbers 5, 6, 9, and 10 corresponding to 1 are the attention area.

FIG. 13 is a diagram showing an example of bit plane division for each tile. As shown in FIG. 13, by including more bit planes of the attention area in the lower layer, while maintaining the accuracy of the area including the human face, the original image 40
It is possible to realize high compression of.

Another embodiment in which the present invention is applied to a still grayscale image will be described with reference to FIG. 6 described above. This example relates to the division of the image area when the user designates the human face portion shown in FIG. 6A as the attention area and wants the bit plane division of the attention area to be different from other areas. It is shown. Here, the unit for area division is a tile. By setting the tile size so as to cover the human face portion of the original image 40, efficient compression control is possible. In the case of this example, high compression of the original image 40 becomes possible by including many bit planes only in the tile number 10 corresponding to the black-painted portion 42 shown in FIG. 6B in the lower layer. Of course, it is not necessary to cover the face portion with one tile, and it may be covered with a plurality of tiles.

In addition, in FIG. 7 described above, the present invention is applied to a still grayscale image, and the area including the human face shown in FIG. 7A is obtained similarly to the example of the tile shown in FIG. In the case where a large number of bit planes are arranged in the lower layer with the area of interest as the attention area, the unit of area division will be described as a precinct. As shown in FIG. 7A, by dividing the original image 40 into a plurality of precincts and including many precinct bit planes corresponding to the black-painted portion 43 shown in FIG. 7B in the lower layer. It is possible to realize high compression of the original image 40 while maintaining the accuracy of the area including the human face.

Further, in FIG. 8 described above, the present invention is applied to a still gray-scale image, and the area including the human face shown in FIG. 8A is obtained similarly to the example of the tile shown in FIG. In the case where a large number of bit planes are arranged in the lower layer with the area of interest as the attention area, the unit of area division will be described as a code block. As shown in FIG. 8 (A),
The original image 40 is divided into a plurality of code blocks as shown in FIG.
By including many bit planes of the code block corresponding to the black-painted portion 44 shown in (B) in the lower layer, the accuracy of the area including the human face can be maintained and the original image 40
It is possible to realize high compression of.

FIG. 14 is a block diagram showing another example of the configuration of an image compression / decompression device for still images according to an embodiment of the present invention. In the figure, 130 is a color space conversion / inverse conversion unit, 131. Is a first component, and the first component 131 is a wavelet transform / inverse transform unit 131.
a, a quantization / inverse quantization unit 131b, a bit plane division unit 131c, and an entropy encoding / decoding unit 131d. Similarly, the second component 132 includes a wavelet transform / inverse transform unit 132a and a quantization / inverse quantization unit. Thirteen
2b, a bit plane division unit 132c, and an entropy coding / decoding unit 132d, and a third component 1
33 is a wavelet transform / inverse transform unit 133a, a quantization / inverse quantization unit 133b, and a bit plane division unit 133.
c, consisting of entropy coding / decoding section 133d, 1
Reference numeral 34 is a code stream processing unit, and 135 is an attention level setting unit. The attention level setting unit 135 in this example sets the attention level for a bit plane obtained by decomposing wavelet coefficients in a divided area such as a tile.

In the image compression / expansion device shown in FIG. 14, the conversion / inverse conversion unit in each processing block is shown as one so as to be capable of compressing and expanding the image. You may take the form divided into parts. By doing so, it becomes possible to independently use the image compression device and the image expansion device. Further, the image compression / decompression device of the present invention, or the compressed image decompressed by the image decompression device is not limited to the compressed image compressed by the image compression device of the present invention, and the image is divided into a plurality of regions,
It is possible to set the attention level for each bit plane of the divided areas, control the compression for each area based on the set attention level, and apply the compression image to all compressed images.

The image compression / decompression device shown above is intended for still images. Processing is component 13
1, 132, 133 are advanced in parallel. The color space here is RGB or YUV.

The case of compression will be described below. First, the color space conversion / inverse conversion unit 130 converts the color space into RGB or YUV color space, and the component 13
The processing is advanced in parallel every 1, 132, 133. Here, the first component 131 will be described as an example, but other second component 132 and third component 13
The same processing is performed for 3. The color space converted image is
It is divided into tiles, and wavelet transform processing is performed by the wavelet transform / inverse transform unit 131a for each tile. Next, the quantization / inverse quantization unit 131b quantizes the wavelet-transformed coefficients.

Next, the bit-plane dividing unit 131c divides the bit-plane consisting of the quantized wavelet coefficients into regions, and when the user sets the attention level to the bit-planes in the region, the set attention is set. The bit plane can be divided and set according to the degree. In this way, the bit plane is divided according to the degree of attention for each area of the image. Furthermore, the wavelet coefficient is
The entropy coding / decoding unit 131d performs coding. Finally, the codestream processing unit 134 generates a codestream. When the code stream is generated, a code string is generated according to the degree of attention as shown in FIG. 18 described later, and the code stream is configured in the order of the degree of attention.

FIG. 15 is a block diagram for explaining the details of another structural example of the image compression / expansion device to which the present invention is applied. In the figure, 140 is a compression / expansion device.
40 is a color space conversion / inverse conversion unit 140a, a two-dimensional wavelet conversion / inverse conversion unit 140b, a quantization / inverse quantization unit 140c, a bit plane division unit 140d, an entropy encoding / decoding unit 140e, and a tag processing unit 140f. , 141 is an image dividing unit, 142 is a compression control unit, 1
Reference numeral 43 is an attention degree setting means, and 144 is a tag processing control means. The compression / decompression device 140 performs the same processing as the processing in each component shown in FIG. Here, the tag processing unit 140f is added as one of the features of JPEG2000, and is a functional block for generating and interpreting a codestream.

The original image has a color space conversion / inverse conversion unit 140a.
After the color space conversion, the two-dimensional wavelet transform is performed, the quantization / inverse quantization unit 140c quantizes it based on the quantization rate, and the image dividing unit 141 divides it into a plurality of regions. The divided areas are called tiles. The quantized wavelet coefficient is divided into layers that are a collection of bit planes in the bit plane division unit 140d. The division is set by the compression control unit 142 according to the attention degree designated by the attention degree setting unit 143.
The compression controlled bit plane is an entropy code.
The decoding unit 140e performs entropy coding, and the tag processing unit 140f processes the code stream. At that time, the tag processing control unit 144 sets the code stream sequence so as to match the degree of attention.

In addition, as a form that is often used,
Two types of bit plane divisions can be set for the divided areas. In this case, an arbitrary target area,
The attention level may be set so that the bit plane division is different from that of other areas. In other words, if a bit plane of the target area is set to have a higher attention level than that set to another area, the image quality of the target area becomes higher in terms of image quality, and the bit plane of the target area is When the attention level that is higher than the attention level set for the area is set to be smaller, the image quality of the target area becomes coarser. In this way, the bit plane division can be set appropriately according to the preference of the user.

According to the present invention, the degree of attention is set for the bit plane over the entire image, and the compression control is performed according to the degree of attention. Therefore, it is possible to realize image compression that suits the individual taste. Further, since the attention level is set for at least two types of regions, it is possible to easily realize the compression control that reflects the user's preference with a simple configuration.

The image compression / decompression device shown in FIG. 15 can be applied to a moving image by adopting the same configuration as the moving image compression / decompression device 30 shown in FIG. At this time, it is possible to set the bit plane division for each area for all images, or to set the method when a change occurs, or to automatically move the target object from the inter-frame difference data etc. It is also possible to use a method of detecting the same and setting the bit plane division.

FIG. 16 is a flow chart for explaining an example of another image compression process to which the present invention is applied. First,
The image compression / decompression device of the present invention receives an input image,
The image is divided into tiles (step S51), it is determined whether color space conversion to RGB, YUV, or the like is performed according to the color space of the input image (step S52), and when it is determined that it is necessary (YES). In the case), color conversion is performed (step S53). In step S52, the input image may be already color space-converted, or it may be compressed without performing the color space conversion. Therefore, in this case (NO), the process proceeds to step S54. next,
Wavelet transformation is performed (step S54) and quantization is performed (step S55).

Next, in step S56, it is determined whether or not the processing has been completed for all areas (step S56). If not completed (in the case of NO),
It is determined whether the attention level is set for each of the divided areas (step S57). When the user sets the attention degree for each bit plane in the area and the division is designated (in the case of YES), the bit plane division is set for the corresponding divided area (step S).
58). If the bit plane division is not designated (NO), the reference bit plane division is set for each area (step S59).

Next, the divided bit planes are encoded (step S60). When the encoding is completed, the counter for counting the divided areas is incremented (step S61), and the process returns to step S56. Hereinafter, steps S56 to S6 are performed until the processing is completed for all areas.
When the process up to 1 is repeated and the processing is completed for all areas (YES in step S56), it is determined again whether the attention level is set (step S62). If the attention level is set, the set attention level is set. In order according to the degree,
A code stream string is generated (step S63) and tag processing is performed. When the attention level is not set, normal code stream generation is performed (step S64), and the process ends.

Here, referring to FIG. 5 described above, a case where the present invention is applied to a still grayscale image will be described by using a tile as a unit for area division. Figure 5
As shown in (A), the original image 40 is divided into a plurality of tiles and, for example, as shown in FIG.
When the tiles are numbered up to, the tiles having tile numbers 5, 6, 9, and 10 corresponding to the black-painted portion 41 shown in FIG. 5B are prioritized. FIG. 17 is a diagram showing an example of attention level setting for each tile according to the present invention, and shows an example of attention level setting between the tiles with tile numbers 5, 6, 9, and 10 and other tiles as priority tiles. . In this way, a high attention level (attention level 0 is the highest level in this example) is set for many bit planes in the priority tiles, and a code string shown in FIG. It is possible to create a code string according to.
Also, due to restrictions on the size of the code string and the like, it is possible to increase the compression rate by cutting the bit plane from the portion of low attention.

FIG. 18 is a diagram showing an example of the structure of a code string when the attention level shown in FIG. 17 is created.
50 is a main header, 151 is a tile part header, 1
52 is a bit stream, and 153 is the end of the code stream. By arranging the code strings in descending order of attention, the most effective image can be obtained even if the data is cut off in the middle due to a failure of the transmission path or the like. Further, for example, it is possible to easily meet the demand that only the data up to the attention level 2 is required. Further, when the decompression device has a predetermined function, it is possible to display the received data in order, so that it is possible to display the data in descending order of attention.

FIG. 19 is a block diagram showing another example of the configuration of an image compression / decompression device for a still image according to an embodiment of the present invention. In the figure, 170 is a color space conversion / inverse conversion unit, 171. Is a first component, and the first component 171 is a wavelet transform / inverse transform unit 171.
a, a quantization / inverse quantization unit 171b, a bit plane division unit 171c, and an entropy encoding / decoding unit 171d. Similarly, the second component 172 includes a wavelet transform / inverse transform unit 172a, a quantization / inverse quantization unit. 17
2b, a bit plane division unit 172c, and an entropy coding / decoding unit 172d.
73 denotes a wavelet transform / inverse transform unit 173a, a quantization / inverse quantization unit 173b, and a bit plane division unit 173.
c, consisting of entropy coding / decoding section 173d, 1
Reference numeral 74 is a code stream processing unit, and 175 is a resolution setting unit.

In the image compression / expansion device shown in FIG. 19, one conversion / inverse conversion unit in each processing block is shown so as to be capable of compressing and expanding an image. You may take the form divided into parts. By doing so, it becomes possible to independently use the image compression device and the image expansion device. Further, the image compression / decompression device of the present invention, or the compressed image decompressed by the image decompression device is not limited to the compressed image compressed by the image compression device of the present invention, and the image is divided into a plurality of regions,
It is possible to set different resolutions, control the compression for each area based on the set resolutions, and apply the compressed images in general.

The image compression / decompression device shown above is intended for still images. Processing is component 17
Every 1,172,173 is advanced in parallel. RGB or YUV can be input as the color space here.

Hereinafter, the case of compression will be described. First, the color space conversion / inverse conversion unit 170 converts the color space into RGB or YUV color space, and the component 17
Processing is advanced in parallel for each of 1,172 and 173. Here, the first component 171 will be described as an example, but other second component 172 and third component 17
The same processing is performed for 3. The color space converted image is
It is divided into tiles, and wavelet transform processing is performed by the wavelet transform / inverse transform unit 171a for each tile.

Next, the quantization / inverse quantization unit 171b quantizes the wavelet-transformed coefficients. At this time, if the resolution is set for each area, the quantization is performed according to the set resolution. Specifically, the high frequency component of the wavelet coefficient is set to 0 or the high frequency component can be cut according to the set resolution. For this, for example, a threshold value is set in advance for the wavelet coefficient value (high frequency component), and the threshold value and the resolution are associated with each other. At this time, when the resolution is set by the user, the wavelet coefficients below the threshold corresponding to the resolution may be cut (the sign is removed), or the wavelet coefficient value below the threshold may be set to zero. Also, as another method, the resolution set by the user and the high frequency components (HL, LH, HH) of the subbands.
Are associated with each other, and the wavelet coefficient of the high frequency component according to the resolution is cut (the sign is removed), or the wavelet coefficient value of the high frequency component is all 0.
You may

Next, in the bit plane division unit 171c,
The bit plane composed of the quantized wavelet coefficients is divided for each area, and the divided wavelet coefficients are further encoded by the entropy encoding / decoding unit 171d. Finally, the code stream processing unit 17
At 4, the codestream is generated.

FIG. 20 is a block diagram for explaining the details of another configuration example of the image compression / decompression device to which the present invention is applied. In the figure, reference numeral 180 denotes a compression / decompression device.
Reference numeral 80 denotes a color space conversion / inverse conversion unit 180a, a two-dimensional wavelet conversion / inverse conversion unit 180b, a quantization / inverse quantization unit 180c, a bit plane division unit 180d, an entropy encoding / decoding unit 180e, and a tag processing unit 180f. 181 is an image dividing unit, 182 is a compression control unit,
Reference numeral 183 is a resolution setting means, and 184 is a tag processing control means. The compression / decompression device 180 performs the same processing as the processing in each component shown in FIG. Here, the tag processing unit 180f is one of the features of JPEG2000.
It is a functional block for generating and interpreting a codestream.

The original image is a color space conversion / inverse conversion unit 180a.
After the color space conversion, the two-dimensional wavelet transform is performed, the quantization / inverse quantization unit 180c performs quantization based on the quantization rate, and then the image dividing unit 181 divides the image into a plurality of regions. The divided areas are called tiles. The quantization is performed according to the resolution set by the resolution setting unit 183. In the present embodiment, the case where the attention area and the non-attention area are divided in tile units has been described, but as another form, division in precinct units, code block units, position (pixel) units, or It is also possible to divide them.

Next, the quantized wavelet coefficient is divided into a layer which is a collection of bit planes in the bit plane dividing unit 180d. The bit plane compression-controlled by the compression control unit 182 according to the resolution set by the resolution setting unit 183 is
Entropy coding is performed by the entropy coding / decoding unit 180e, and codestream processing is performed by the tag processing unit 180f. At this time, the tag processing control unit 184 sets the code stream string so as to match the resolution.

In addition, as a form often used,
At least two types of resolutions can be set for the divided areas. In this case, the resolution may be set so that the arbitrary target area and the other areas have different resolutions. That is, when a higher resolution is set for the target area than the resolution set for other areas, the target area has higher definition in terms of image quality, and the target area has a higher resolution than the resolution set for other areas. If you also set a lower resolution,
The image quality of the target area is coarser. In this way, the resolution can be set appropriately according to the preference of the user.

Further, in the configuration of this embodiment, the calculation of the wavelet coefficient can be appropriately omitted according to the resolution set by the resolution setting means 183.
For this purpose, for example, the calculation is omitted for the high frequency components (LH, HL, HH) of the wavelet coefficients. As a result, it is possible to speed up the processing by omitting extra calculation. Furthermore, to calculate the low frequency component of the wavelet coefficient in a specific region, use an average calculation,
A method of extracting and using a specific element from the low-frequency component is also effective in increasing the processing speed.

According to the present invention, the resolution is set for each of the divided areas, and the compression control is performed according to this resolution. Therefore, it is possible to finely control the image quality for each area. In addition, since the resolution is set for at least two types of areas, it is possible to easily realize compression control that reflects the user's preference with a simple configuration. Further, by appropriately omitting the calculation of the wavelet coefficient, it is possible to omit an extra process and speed up the process.

The image compression / decompression device shown in FIG. 20 can be applied to a moving image by adopting the same configuration as the moving image compression / decompression device 30 shown in FIG. At this time, it is possible to set the bit plane division for each area for all images, or to set the method when a change occurs, or to automatically move the target object from the inter-frame difference data etc. It is also possible to use a method of detecting the same and setting the bit plane division.

FIG. 21 is a flow chart for explaining an example of another image compression process to which the present invention is applied. First,
The image compression / decompression device of the present invention receives an input image,
The image is divided into tiles (step S71), it is determined whether color space conversion to RGB, YUV, or the like is performed according to the color space of the input image (step S72), and when it is determined that it is necessary (YES). In the case), color conversion is performed (step S73). In step S72, it is natural that the input image is already color space-converted, or the image is compressed without color space conversion. In this case (NO), the process proceeds to step S74. next,
Wavelet transform is applied (step S74).

Next, in step S75, it is determined whether or not the processing has been completed for all areas, and if not completed (in the case of NO), it is determined whether or not the resolution has been set for each area. (Step S7
6). If the resolution is set (YES), the corresponding region is quantized in accordance with the resolution (step S77). If the resolution is not set (NO), the reference resolution is set for each area and quantization is performed (step S78).

Next, the relevant area is divided into bit planes (step S79) and encoded (step S8).
0) is performed. When the encoding is completed, the counter for counting the divided areas is incremented (step S8).
1) and returns to step S75. Hereinafter, the processing from step S75 to step S80 is repeated until the processing is completed for all areas, and when the processing is completed for all areas (step S
In the case of YES at 75), the code stream sequence is generated (step S82), and the process is ended.

Here, referring to FIG. 5 described above, an example of tile division of an area when the present invention is applied to a still grayscale image and an example of area division when a resolution is set will be described. When the original image 40 is divided into a plurality of tiles as shown in FIG. 5A and tiles are numbered from 0 to 15 in raster order as shown in FIG. 25, for example, in FIG. The tiles with tile numbers 5, 6, 9, and 10 corresponding to the black-painted portion 41 shown are set as priority targets (target areas). At this time, for example, when the resolution of the attention area 41 is the same as that of the original image 40 and the resolution of the non-attention area is ¼ of the resolution of the attention area 41, components other than the LL component of the wavelet coefficient ( High frequency component: L
H, HL, HH) are all set to 0 or there is no sign.

According to the present invention, high resolution can be realized while maintaining high image quality by finely setting the resolution for the attention area and the non-attention area and performing compression according to the set resolution. In addition to this, it is possible to reduce the processing time at that time.

Although the respective embodiments have been described above centering on the image compressing apparatus, the image decompressing apparatus, the image compressing method, and the image decompressing method of the present invention, the present invention can be applied to a computer as these apparatuses or of these apparatuses. It may be in the form of a program for causing each means to function, or for causing a computer to execute these methods, or as a computer-readable recording medium recording the program.

An embodiment of a recording medium storing a program and data for realizing the image compression and decompression functions according to the present invention will be described. As the recording medium, specifically,
CD-ROM, magneto-optical disk, DVD-ROM, floppy (registered trademark) disk, flash memory, memory card, memory stick and other various ROMs and Rs
An AM or the like can be assumed, and by causing a computer to execute the functions according to the above-described embodiments of the present invention on these recording media, and recording and distributing a program for realizing the image compression and decompression functions, Facilitate the realization of. Then, the recording medium as described above is attached to an information processing device such as a computer and the program is read by the information processing device, or the program is stored in a storage medium included in the information processing device, and as necessary. By reading out, the image compression and decompression functions according to the present invention can be executed.

According to the present invention, a program having a function of an image compression device or an image expansion device can be executed, and by executing the program, it is possible to realize an image compression / expansion process that suits an individual's taste. .

[0175]

According to the present invention, an important region of an image and a region of interest can be set by the user of the image, and the set region can be image-compressed with the accuracy required by the user. Accordingly, it is possible to provide an image compression device, an image decompression device, an image compression / decompression device, a computer-executable program, and a recording medium that records the program, which can be easily realized.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration example of an image compression / decompression device for a still image according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating details of a configuration example of an image compression / decompression device to which the present invention is applied.

FIG. 3 is a block diagram showing a configuration example of an image compression / decompression device for a moving image according to the present invention.

FIG. 4 is a flowchart illustrating an example of image compression processing to which the present invention is applied.

FIG. 5 is a diagram showing an embodiment in which the present invention is applied to a still grayscale image.

FIG. 6 is a diagram showing another embodiment in which the present invention is applied to a still grayscale image.

FIG. 7 is a diagram showing an example in which a precinct is applied to area division.

FIG. 8 is a diagram showing an example in which a code block is applied to area division.

FIG. 9 is a flowchart illustrating an image decompression method according to an exemplary embodiment of the present invention.

FIG. 10 is a block diagram showing another configuration example of an image compression / decompression device for a still image according to an embodiment of the present invention.

FIG. 11 is a block diagram for illustrating details of another configuration example of the image compression / decompression device to which the present invention is applied.

FIG. 12 is a flowchart illustrating an example of another image compression process to which the present invention is applied.

FIG. 13 is a diagram showing an example of bit plane division for each tile.

FIG. 14 is a block diagram showing another configuration example of an image compression / decompression device for still images according to an embodiment of the present invention.

FIG. 15 is a block diagram for illustrating details of another configuration example of the image compression / decompression device to which the present invention is applied.

FIG. 16 is a flowchart illustrating an example of another image compression process to which the present invention is applied.

FIG. 17 is a diagram showing an example of attention level setting for each tile according to the present invention.

FIG. 18 is a diagram showing a configuration example of a code string when the attention level shown in FIG. 17 is created.

FIG. 19 is a block diagram showing another configuration example of an image compression / decompression device for a still image according to an embodiment of the present invention.

FIG. 20 is a block diagram illustrating details of another configuration example of the image compression / decompression device to which the present invention is applied.

FIG. 21 is a flowchart illustrating an example of another image compression process to which the present invention is applied.

FIG. 22 is a block diagram for explaining the basics of the JPEG algorithm.

FIG. 23 is a block diagram for explaining the basics of the JPEG2000 algorithm.

FIG. 24 is a diagram showing an example of subbands at each decomposition level when the decomposition level is 3.

FIG. 25 is a diagram illustrating an example of each component of a tile-divided color image.

FIG. 26 is a diagram illustrating an example of a relationship between precincts and code blocks.

FIG. 27 is a diagram briefly showing an example of the structure of a codestream.

FIG. 28 is a diagram briefly showing an example of a procedure of ranking bit planes.

[Explanation of symbols]

10, 20a, 50, 60, 110, 120a, 13
0,140a, 170,180a ... Color space conversion / inverse conversion
Part, 11, 111, 131, 171 ... First component
G, 12, 112, 132, 172 ... Second component
G, 13, 113, 133, 173 ... Third component
11a, 12a, 13a, 20b, 61, 111
a, 112a, 113a, 120b, 131a, 132
a, 133a, 140b, 171a, 172a, 173
a, 180b ... Wavelet transform / inverse transform unit, 11
b, 12b, 13b ... Quantization rate selection unit, 11c, 12
c, 13c, 20c, 52, 62, 111b, 112
b, 113b, 120c, 131b, 132b, 133
b, 140c, 171b, 172b, 173b, 180
c ... Quantization / inverse quantization unit, 11d, 12d, 13d, 2
0d, 53, 63, 111d, 112d, 113d, 1
20e, 131d, 132d, 133d, 140e, 1
71d, 172d, 173d, 180e ... Entropy
Coding / decoding unit, 14, 114, 134, 174 ... Co
Stream processing unit, 15 ... Quantization rate setting unit, 20, 1.
20, 140, 180 ... Compression / expansion device, 20e, 64,
120f, 140f, 180f ... Tag processing unit 21, 1
21, 141, 181 ... Image dividing means, 22, 142,
182 ... Compression control means, 23 ... Quantization rate setting means, 2
4, 124 ... Area designation means, 30 ... Moving image compression / decompression device
, 30a ... Frame control means, 30b ... Still image compression
Decompression means, 30c ... Compressed moving image processing means, 40, 90,
160 ... Original image, 41, 42, 43, 44 ... Region of interest
(Important area), 51 ... Discrete cosine transform / inverse transform unit, 7
0 ... Decomposition Level 0 tile, 71 ... Decomposition
Position level 1 tiles, 72 ... Decomposition
Level 2 tiles, 73 ... Decomposition level 3
Tile, 80 ... First component (R), 81 ... Second
Component (G), 82 ... Third component
(B), 80_t, 81_t, 82 _t, 90_t0, 90_t1, 90
_t2, 90_t3, 160_t0, 160_t1, 160_t2, 160_{t 3}
… Tiles, 90_p4, 90_p6, 160_p0, 160_p1, 16
0_p2, 160_p3… Precinct, 90_b1, 160_b3, 1
60_w3… Code blocks, 100, 150… Main head
Da, 101, 151 ... Tile part header, 102, 1
52 ... bit stream, 103, 153 ... code list
Ream end, 111c, 112c, 113c, 120
d, 131c, 132c, 133c, 140d, 171
c, 172c, 173c, 180d ... Bit plane equivalent
Split part, 115 ... Bit plane setting part, 122 ... Bit
Plane division control means, 123 ... Bit plane division finger
Setting means, 135 ... Attention degree setting unit, 143 ... Attention degree setting hand
Stage, 144, 184 ... Tag processing control means, 175 ... Resolution
Degree setting section 183 ... Resolution setting means.

Continued front page    F-term (reference) 5C059 KK03 KK04 MA00 MA24 MC11                       MC38 ME01 PP01 PP04 PP15                       PP16 SS15 SS21 SS28 TA46                       TB08 TC34 UA02 UA05                 5C078 BA53 CA02 CA27 DA01 DA11                       DB00                 5J064 AA01 BA09 BA16 BC01 BC02                       BC08 BC14 BC16 BC29 BD03

Claims (57)

[Claims] 1. An image dividing means for dividing an image into a plurality of areas, an area specifying means for specifying one or a plurality of attention areas in the image, and a quantization rate for each area divided by the image dividing means. Quantization rate setting means for controlling the compression for each area based on the set quantization rate, and a quantum different from other areas for the attention area specified by the area specifying means. An image compression device characterized in that a conversion rate can be set. 2. The image compression apparatus according to claim 1, further comprising: attention level setting means for setting the attention level to the attention area designated by the area designation means, wherein the attention area to which the attention degree is set is set. On the other hand, the image compression device is characterized in that different quantization rates are set according to the degree of attention. 3. The image compression apparatus according to claim 1, wherein at least two different quantization rates are set for the areas divided by the image dividing means. 4. The image compression apparatus according to claim 3, wherein the quantization rate of the attention area designated by the area designating unit is set higher than that of other areas. 5. The image compression apparatus according to claim 3, wherein the quantization rate of the attention area designated by the area designation unit is set to be lower than that of other areas. 6. An image dividing means for dividing an image into a plurality of areas, an area specifying means for specifying one or a plurality of attention areas in the image, and a bit number for each area divided by the image dividing means. It has a bit number setting means and a compression control means for controlling compression for each area based on the set bit number, and sets a bit number different from other areas for the attention area specified by the area specifying means. An image compression device capable of enabling. 7. The image compression apparatus according to claim 6, further comprising: attention level setting means for setting the attention level to the attention area designated by the area designation means, wherein the attention area to which the attention degree is set is set. On the other hand, an image compression device characterized in that a different number of bits is set according to the degree of attention. 8. The image compression apparatus according to claim 6, wherein at least two different bit numbers are set in the area divided by the image dividing means according to the attention level. Compressor. 9. The image compression apparatus according to claim 8, wherein the number of bits of the attention area designated by the area designating unit is set to be larger with respect to other areas as the attention degree is higher. Image compression device. 10. The image compression apparatus according to claim 8, wherein the number of bits of the attention area designated by the area designating unit is set to be smaller with respect to other areas as the attention degree is higher. Image compression device. 11. The image compression apparatus according to claim 1, wherein the area designating unit designates one or a plurality of regions of interest from among the regions divided by the image dividing unit. An image compression device characterized by. 12. The image compression apparatus according to claim 1, wherein the area designating unit designates one or a plurality of areas of interest in accordance with a portion of interest in the image. Image compression device. 13. An image dividing means for dividing an image into a plurality of areas, a resolution setting means for setting a resolution for each area divided by the image dividing means, and a compression with a different resolution for each area according to the set resolution. An image compression apparatus comprising: a compression control unit for controlling. 14. The image compression apparatus according to claim 13, wherein the area divided by the image dividing unit is
An image compression apparatus characterized by setting at least two different resolutions. 15. The image compression apparatus according to claim 14, wherein the area divided by the image dividing unit is
An image compression apparatus, characterized in that the resolution of at least one area is set higher than that of other areas. 16. The image compression apparatus according to claim 14, wherein the area divided by the image dividing unit is
An image compression apparatus, wherein the resolution of at least one area is set lower than that of other areas. 17. The image compression apparatus according to claim 13, wherein calculation of wavelet coefficients is omitted according to the resolution set by the resolution setting unit. . 18. The image compression apparatus according to claim 17, wherein the wavelet coefficient for which the calculation is omitted is a high frequency component. 19. The image compression apparatus according to claim 18, wherein an average calculation is used to calculate the low-frequency component of the wavelet coefficient of the specific region. 20. The image compression apparatus according to claim 18, wherein a specific element is extracted and used for the calculation of the low frequency component of the wavelet coefficient of the specific region. 21. The image compression apparatus according to claim 1, wherein tiles are used to divide the image into a plurality of areas. 22. The image compression apparatus according to claim 1, wherein a precinct is used to divide an image into a plurality of areas. 23. The image compression apparatus according to claim 1, wherein a code block is used to divide an image into a plurality of areas. 24. The image compression apparatus according to claim 1, wherein when the image is divided into a plurality of areas, the image is divided in pixel units. 25. The image compression apparatus according to claim 1, wherein when the image is divided into a plurality of areas, any two or more of a tile, a precinct, a code block, and a pixel unit are combined. An image compression device characterized in that it is divided into two parts. 26. The image compression apparatus according to claim 1, wherein when the image is divided into a plurality of areas, the area is set according to a target portion in the image, and the set area is set. An image compression apparatus, characterized in that an image is divided by. 27. The image compression apparatus according to claim 26, wherein the boundaries of the divided areas are tile boundaries. 28. The image compression apparatus according to claim 26, wherein the boundary of the divided areas is a precinct boundary. 29. The image compression apparatus according to claim 26, wherein a boundary between the divided areas is a code block boundary. 30. The image compression apparatus according to claim 26, wherein the boundaries of the divided areas are two or more of a tile boundary, a precinct boundary, and a code block boundary. apparatus. 31. An image dividing unit for dividing an image into a plurality of regions, a bit plane decomposition unit for dividing the region divided by the image dividing unit into bit planes, and a degree of attention is set for the decomposed bit planes. And a compression control unit that controls compression for each area based on the attention level set for the bit plane, and generates a compressed image according to the attention level set for the bit plane. An image compression device characterized by the above. 32. An image dividing unit that divides an image into a plurality of regions, a bit plane decomposition unit that decomposes the region divided by the image dividing unit into bit planes, and a degree of interest is set for the decomposed bit planes. An attention degree setting means, a compression control means for controlling compression for each area based on the attention degree set in the bit plane, and a tag processing control means for controlling tag processing based on the attention degree. An image compression apparatus, which generates a compressed image according to the degree of attention set in the bit plane. 33. The image compression apparatus according to claim 31, wherein the attention level is set for a bit plane corresponding to each area divided by the image division means. 34. The image compression apparatus according to claim 33, wherein at least two different attention levels are set for bit planes corresponding to respective regions divided by the image dividing means. Compressor. 35. The image compression apparatus according to claim 33, wherein at least one of the image data is divided by the image division means.
An image compression apparatus characterized in that, in a bit plane of one image area, a high degree of attention is set with respect to the degree of attention set in another image area. 36. The image compression apparatus according to claim 33, wherein at least one image is divided by the image dividing means.
An image compression apparatus characterized in that, in a bit plane of one image area, a high attention degree is set to be smaller than a attention degree set to another image area. 37. The image compression apparatus according to claim 1, wherein the image compression apparatus is configured by a combination of a two-dimensional discrete wavelet transformer, a quantizer, and an entropy encoder. Image compression device. 38. An image decompression device for decompressing a compressed image compressed by using the image compression device according to claim 31, wherein the compressed image is included in the compressed image. An image decompression device, which decompresses based on information regarding degree. 39. An image decompression device for decompressing a compressed image compressed by using the image compression device according to claim 13, wherein the resolution is included in the compressed image. An image decompression device, which decompresses based on information about the image. 40. An image decompression device for decompressing a compressed image compressed by using the image compression device according to any one of claims 1 to 37, wherein the compressed image is included in the compressed image. An image decompressing device, which decompresses based on information of an area divided by a dividing means. 41. An image is divided into a plurality of areas, a quantization rate is set for each of the divided areas, compression is controlled for each area based on the set quantization rate, and a compressed image compressed is obtained. An image decompressing device for decompressing, wherein the compressed image is decompressed based on information of divided areas included in the compressed image. 42. An image is divided into a plurality of areas, the number of bits is set for each of the divided areas, compression is controlled for each area based on the set number of bits, and a compressed compressed image is expanded. An image decompression device, which decompresses a compressed image based on information of a divided area included in the compressed image. 43. The image decompression device according to claim 38, which is configured by a combination of a two-dimensional discrete wavelet inverse transformer, an inverse quantizer, and an entropy decoder. An image decompression device. 44. An image is divided into a plurality of areas, a quantization rate is set for each of the divided areas, compression is controlled and compressed for each area based on the set quantization rate, and the compression is performed. An image compression / decompression device for decompressing a compressed image, which comprises a combination of a two-dimensional discrete wavelet transform / inverse transformer, a quantizer / inverse quantizer, and an entropy encoder / decoder. Characteristic image compression / decompression device. 45. An image is divided into a plurality of areas, the number of bits is set for each of the divided areas, compression is controlled for each area based on the set number of bits, and the compressed image is compressed. Is an image compression / decompression device for decompressing, which is configured by a combination of a two-dimensional discrete wavelet transform / inverse transformer, a quantizer / inverse quantizer, and an entropy encoder / decoder. Image compression and expansion device. 46. An image is divided into a plurality of areas, the divided areas are decomposed into bit planes, an attention degree is set for the decomposed bit planes, and compression is performed for each area based on the set attention degree. An image compression / decompression device for controlling, controlling tag information based on the set degree of attention to generate a compressed image, and decompressing the generated compressed image, comprising a two-dimensional discrete wavelet transform / inverse converter, An image compression / decompression device comprising a combination of a quantizer / inverse quantizer and an entropy encoder / decoder. 47. A means for dividing an image into a plurality of regions,
Means for setting a resolution for each of the divided areas, means for controlling compression at a different resolution for each area according to the set resolution, and expanding the compressed compressed image based on the resolution set for the compressed image An image compression / decompression device, comprising: 48. The image compression apparatus according to claim 1, further comprising a two-dimensional discrete wavelet transform / inverse transformer, a quantizer / inverse quantizer, and an entropy encoder / decoder. An image compression / decompression device characterized by being configured in combination with. 49. An image division step of dividing an image into a plurality of areas, an area designation step of designating one or a plurality of attention areas in the image, and a quantization rate for setting a quantization rate for each of the divided areas. A setting step and a compression control step of controlling compression for each area based on the set quantization rate, and it is possible to set a quantization rate different from other areas for the attention area. Image compression method. 50. An image dividing step of dividing an image into a plurality of areas, an area specifying step of specifying one or a plurality of attention areas in the image, and a bit number setting step of setting a bit number for each of the divided areas. And a compression control step of controlling compression for each area based on the set number of bits, and it is possible to set a different number of bits for the attention area from other areas. . 51. An image division step of dividing an image into a plurality of areas, a bit plane decomposition step of decomposing the divided area into bit planes, and an attention degree setting for setting an attention degree to the decomposed bit planes. The bit plane includes a step, a compression control step that controls compression for each area based on the attention level set in the bit plane, and a tag processing control step that controls tag processing based on the attention level. An image compression method characterized in that a compressed image is generated according to a set attention level. 52. An image division step of dividing an image into a plurality of areas, a resolution setting step of setting a resolution for each of the divided areas, and a compression control for controlling compression at a different resolution for each area according to the set resolution. An image compression method, comprising: 53. An image is divided into a plurality of areas, a quantization rate is set for each of the divided areas, compression is controlled for each area based on the set quantization rate, and a compressed image compressed is obtained. An image decompression method for decompressing, which comprises decompressing a compressed image based on information of a divided area included in the compressed image. 54. An image is divided into a plurality of areas, the number of bits is set for each of the divided areas, compression is controlled for each area based on the set number of bits, and a compressed compressed image is expanded. An image decompression method, which comprises decompressing a compressed image based on information on divided areas included in the compressed image. 55. An image is divided into a plurality of areas, a degree of attention is set for each bit plane of the divided areas, compression is controlled for each area based on the set degree of attention, and the set degree of attention is set. An image decompression method for decompressing a compressed image generated by controlling tag processing on the basis of the above, wherein the compressed image is decompressed based on the information regarding the degree of attention included in the compressed image. Image decompression method. 56. The image compression apparatus according to any one of claims 1 to 37 or each means of the image compression apparatus,
Alternatively, a program for causing a computer to function as the image decompression device according to any one of claims 38 to 43 or each unit of the image decompression device. 57. The image compression apparatus according to any one of claims 1 to 37 or each means of the image compression apparatus,
Alternatively, a computer-readable recording medium having recorded therein a program for causing a computer to function as the image decompression device according to any one of claims 38 to 43 or each unit of the image decompression device.