JP2000333019A - Image compression apparatus and method, and printer apparatus provided with the image compression apparatus - Google Patents

Abstract

(57) [Summary] [Problem] For both a photograph image and a text image, and an image in which a photograph and a text are mixed,
Achieve high compression ratio without deteriorating image quality. An image is divided into partial images of a predetermined size,
By performing image compression on the divided partial images based on a plurality of image compression algorithms and selecting a compression result having the best compression rate as the compression result of the partial image, independent image compression is performed for each partial image. Performs image compression using an algorithm.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image compression technique.
In particular, it is used for data handled by printers, copiers, scanners, facsimiles, displays, etc., and can achieve a high compression ratio without deteriorating image quality, regardless of the type of image such as a photograph or text. The present invention relates to an image compression technique.

2. Description of the Related Art In a device such as a printer, it is required to hold a large number of input images. Therefore, in general, an image is compressed in the device. Algorithms for performing such image compression include, for example, those based on the JPEG scheme and those based on the palette compression scheme.

An image compression algorithm based on the JPEG method is based on a method of irreversibly encoding based on DCT (discrete cosine transform). Specifically, the input image is divided into blocks of 8 × 8 pixels, and the DCT operation is performed for each block. Then, a DC component (DC component) and an AC component (AC component) are independently quantized with respect to the obtained DCT coefficient, and the difference between the DC component coefficient and the DC component coefficient of the immediately preceding block is used as a predicted value. , And Huffman coding is performed on the coefficient of the AC component.

[0003] The JPEG method is suitable for images input by photographs or scanners because it can cope with many color spaces, and can achieve a particularly high compression ratio when the change in image density is small. However, the compression ratio is deteriorated for an image having a sharp outline such as a text image, and since a high-frequency component is lost in the process of compression / expansion, the periphery of the outline ( It has the characteristic that the image quality of the edge portion is deteriorated.

The image compression algorithm based on the pallet compression method is a reversible compression algorithm that can be applied when the number of density values in an image is equal to or less than a predetermined number. Specifically, pallets are prepared for the number of types of density values in the image, density values are stored in each pallet, and pallet numbers are stored instead of density values for each pixel.

[0005] Since the palette compression method is lossless compression,
Even if the image has a sharp outline such as a text image, the image quality does not deteriorate. However, if the number of types of densities in an image exceeds a predetermined number, it cannot be applied as it is, so that it generally cannot exhibit sufficient performance for images such as photographs having many density values. have.

In a printer or the like, it is necessary to input and output a wide variety of images from a photograph to a text. However, there has been no image compression technique suitable for both a photographic image and a text image. .

That is, the JPEG system is suitable for compressing a photographic image, but is not suitable for compressing a text image.
Conversely, the pallet compression method is suitable for compressing text images, but is not suitable for compressing photographic images and the like.

In such a case, it is possible for a person to judge the type of image and to determine which image compression technique to use, so that appropriate image compression according to the image can be performed. When characters and figures are mixed in an image, the determination becomes very difficult. Also, JP
Regardless of which of the EG method and the pallet compression method is used, there is a possibility that sufficient results cannot be obtained in terms of image quality and compression ratio.

On the other hand, a printer or the like has an image memory of a fixed capacity inside the apparatus. The compressed image is stored on the image memory, but depending on the content of the original image, sufficient compression cannot be performed even if the image is compressed, and the image memory may not be able to store the image completely. there were.

In order to always be able to store the data in the image memory, the compression ratio should always be equal to or less than a fixed value. However, depending on the image compression algorithm, the compression ratio changes depending on the contents of the image, so that the image quality does not deteriorate. However, it was difficult to keep the compression ratio below a certain value. In particular, the JPEG method has a problem that the memory is likely to be overfilled because the compression ratio of an image containing many high-frequency components, such as an image of a beach, grass, densely populated leaves, or an image with a lot of noise, is poor.

The present invention is directed to an image processing apparatus which can achieve a high compression ratio without deteriorating the image quality for both a photographic image or the like and a text image, or for an image in which a photographic image or the like is mixed. It is intended to provide a compression technique.

[0011] Further, the present invention provides an image processing method that does not depend on the content of an image.
An object of the present invention is to always keep the compression ratio below a certain value without deteriorating the image quality.

SUMMARY OF THE INVENTION The present invention divides an image into partial images of a predetermined size, and performs image compression on the divided partial images based on a plurality of image compression algorithms. By selecting a compression result having a high rate as a compression result of the partial image, image compression is performed by an independent image compression algorithm for each partial image.

The image compression method of the present invention can be implemented by a computer, and a computer program for the method is installed or loaded on the computer through various media such as a CD-ROM, a magnetic disk, a semiconductor memory, and a communication network. Can be.

The present invention also provides image dividing means for dividing an image into partial images of a predetermined size, and performs image compression on the divided partial images based on a plurality of image compression algorithms. A compression result storage unit that stores a compression rate and a compression result in association with each of the plurality of image compression algorithms for each of the divided partial images, and by referring to the compression result storage unit, for each of the divided partial images, Compression result selection means for selecting a compression result corresponding to an image compression algorithm having a good compression ratio and outputting the selected result together with identification information of the image compression algorithm.

The partial image having the predetermined size is a partial image of 8 × 8 pixels. The plurality of image compression algorithms are at least an algorithm based on the JPEG system, an algorithm based on the pallet compression system, and only upper bit information for each pixel. It is preferable to include an algorithm to leave.

Further, when the compression ratios of the plurality of image compression algorithms are approximately the same, it is preferable to preferentially select a compression result corresponding to the image compression algorithm that is lossless compression.

Further, the plurality of image compression algorithms include an algorithm based on a JPEG system, an algorithm based on a pallet compression system, and an algorithm for retaining only upper bit information for each pixel. If they are almost the same, the algorithm using the pallet compression method, JPE
It is preferable to preferentially select the corresponding compression result in the order of the algorithm according to the G method and the algorithm that leaves only the upper bit information for each pixel.

Further, according to the present invention, an algorithm used for image compression is identified based on identification information of the image compression algorithm for each of the divided partial images, and an image corresponding to the identified image compression algorithm is identified. It is preferable to include an image decoding unit that performs image expansion based on an expansion algorithm.

According to the present invention having the above structure, a plurality of image compression algorithms are applied to each of the divided partial images,
Since the compression result based on the algorithm with the best compression ratio is stored as the compression result of the partial image, the image is automatically converted without the human being judging the type of image such as a photo image or a text image and selecting an algorithm. An appropriate image compression algorithm is selected according to the type.

Further, since the compression result having the highest compression ratio is independently selected for each partial image, even if the image includes characters, graphics, etc. in the photograph or the like, the photograph or the like is used in the portion of the photograph or the like. Compression suitable for the character is automatically selected for characters and graphic parts.

Further, as a plurality of image compression algorithms, a compression result obtained by an algorithm that leaves only upper bit information for each pixel can be selected, and the compression ratio of such an algorithm is determined independently of the content of the image. , The compression ratio can always be kept below a certain value.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described with reference to FIG. An image compression apparatus 1 according to a first embodiment of the present invention includes an image dividing unit 10, a compression result storage unit 11, and a compression result selection unit 12.

Now, it is assumed that the input image is a grayscale image, and the depth of each pixel is 8 bits, that is, the density value is 256 gradations. Note that the input image may be a color image, in which case the present invention may be applied to each plane such as RGB.

The image dividing means 10 divides the input image into, for example, a partial image of a block of 8 × 8 pixels, and outputs it to the compression result storing means 11. In the present embodiment, description will be made assuming that partial images are processed one by one. However, since processing for each partial image is independent, it may be performed in parallel. Note that the size of the block may be larger than 8 × 8 pixels, and may be determined based on the content of the image, for example, the size of a character or the like in the image.

The compression result storage means 11 receives a partial image as an input, and uses a plurality of image compression algorithms for each partial image, for example, a JPEG system, a palette compression system, and a compression system (hereinafter, referred to as an algorithm) that leaves only upper bit information for each pixel. The image compression is performed by three image compression algorithms of "compulsory compression method". FIG. 2 is a diagram showing an example of the configuration of the compression result storage means 11.
A compression unit 110, a pallet compression unit 111, and a forced compression unit 112 are provided.

The JPEG compression means 110 performs image compression on the input partial image by the JPEG method, and stores a compression ratio and a compression result. Partial image size is 8
If it is larger than a block of × 8 pixels, it is further divided to a size of 8 × 8 pixels and then compressed.

As a specific procedure, first, a frequency distribution is obtained by DCT for a block of 8 × 8 pixels. FIG. 3A shows an example of a partial image, and FIG. 3B shows an example of a frequency distribution obtained by DCT. Next, quantization is performed by dividing each element of the frequency distribution by each element of a predetermined quantization matrix. FIG. 3C shows an example after quantization. Here, the value of the DC component (the upper left element) is directly stored. AC components (elements other than the upper left) are zigzag-scanned and arranged in a row, run-length compressed, Huffman-encoded, and the encoded sequence is stored. Note that run-length compression and Huffman coding are well-known techniques in the field of image compression. For example, “Introduction to Illustrated Digital Image Processing”, Sanno University Press, 1996
Is described in

The JPEG compression means 110 adds the number of bits required to store the DC component and the number of bits required to store the coded sequence, and calculates the density of the original block of 8 × 8 pixels. 5 needed to memorize the value
The compression ratio is obtained by dividing by 12 bits and stored.

The pallet compressing means 111 scans the input partial image and calculates and stores the number of types of density values in the partial image, that is, how many types of density values appear in the partial image.

The pallet compression means 111 performs pallet compression only when the number of types of density values in the partial image is equal to or less than a predetermined number (for example, 8 types or less), and the number of types of density values exceeds the predetermined number. If not, pallet compression is not performed.
In such a case, the compression result is not stored, and the compression ratio is stored as 100%. However, when the contents of the partial image are mainly characters, figures, and the like, the types of density values in the partial image are sufficiently reduced, and it can be expected that pallet compression is performed.

Next, each density value appearing in the partial image is stored in each palette. Now, since the depth of the pixel is 8 bits, the palette is also composed of 8 bits. Pallets are prepared for the number of types of density values, and each pallet is assigned a number for specifying the pallet. For example, when the number of types of density values is eight, 0 to 7
Up to the number.

Then, a pallet number corresponding to each pixel is stored. The number of bits required to store the pallet number is 0 if the number of density values is one,
1 bit for types, 2 bits for 3-4 types, 5 bits
If there are eight types, three bits are used.

FIG. 4 is a schematic diagram showing the result of compression by the pallet compression method. As shown in FIG. 4, the compression result includes information on the number of types of density values, the palette, and the palette number of each pixel. When the number of types of density values is eight, the number of bits required after compression is indicated. Three bits are used to store the number of types of density values, and 64 bits and 64 pixels are used to store eight types of palettes. 3 bits × 64 pixels = 192 bits to store the pallet number of the minute, which is 259 bits in total. Therefore, the compression ratio is 259
/ 512, which is about 50.6%. The compression ratio thus obtained is stored together with the compression result.

In the case of the pallet compression method, since the compression ratio is determined by the number of types of density values, the correspondence between the number of types of density values and the compression ratio can be obtained in advance. FIG. 5 shows the correspondence between the number of types of density values and the compression ratio when the size of the partial image is 8 × 8 pixels.

Therefore, for example, the pallet compressing means 111
When the number of types of density values in the partial image is calculated, the corresponding compression ratio is obtained and stored, and the compression process is performed when the compression result selection unit 12 described later selects the pallet compression method. You may continue and obtain | require a compression result.

In the above-described operation of the pallet compressing means 111, the partial image is scanned a plurality of times. However, when a new density value appears during the scanning of the partial image, the density values are sequentially stored in the pallet. Is stored, and the pallet number corresponding to each pixel is stored while scanning, so that the compression process can be performed by one scan.

The forcible compression means 112 scans the input partial image, extracts only the upper n bits from the 8 bits representing the density value of each pixel, and stores them for each pixel. Here, the value of n may be determined according to the design.
When the value of n is reduced, the compression ratio is improved, but the image quality is reduced. However, in a printer or the like, since the image quality is reduced to some extent by halftone processing used for printing, even if the value of n is set to about half the number of bits per pixel in the input image (for example,
If the depth of one pixel is 8 bits, even if n = 4), printing can be performed to such an extent that deterioration of image quality cannot be judged by human eyes.

In the case of the forced compression method, a fixed value (n / 8) is stored as the compression ratio because the compression ratio is directly determined by the value of n and does not depend on the content of the partial image. For example,
When n = 4, the compression ratio is 50%. This corresponds to multiplying the density value of each pixel by 1/16.

The compression result selection means 12 refers to the compression result storage means 11 to read and compare the compression ratios corresponding to a plurality of image compression algorithms for each partial image. When the compression result storage means 11 has a configuration as shown in FIG. 2, the compression result selection means 12 outputs the JPEG compression means 110 and the pallet compression means 11 for each partial image.
1. The compression ratio is read from the forced compression means 112 and compared.

Next, the compression result corresponding to the image compression algorithm having the best compression rate is selected as the compression result of the partial image. Here, a good compression ratio means that the value of the compression ratio is 0%.
Means close to.

When the compression ratios are approximately the same (for example, when the difference between the compression ratios is approximately 1%), based on the priorities set in advance for the image compression algorithm, Select the compression result. The priority may be set so that lossless compression is prioritized over lossy compression. For example, when the compression result storage means 11 has a configuration as shown in FIG. 2, the priority is set in the order of the palette compression method, the JPEG compression method, and the forced compression method.

Then, the selected compression result is stored as a compressed partial image for the partial image together with the identification information of the image compression algorithm having the best compression rate. Here, the identification information of the image compression algorithm is information for identifying which image compression algorithm has been used for compression. For example, if the compression result storage means 11 has a configuration as shown in FIG. 2, there are three types of image compression algorithms.
, The identification information can be composed of 2 bits.

The compression result selecting means 12 sets a set of compressed partial images for all the partial images and outputs the set as a compressed image for the input image. FIG. 6 is a schematic diagram showing data of a compressed image.

Next, a second embodiment of the present invention will be described with reference to FIG. A printer device 2 according to a second embodiment of the present invention includes an image input unit 20, an image compression device 2
1, a compressed image storage means 22, an image decoding means 23, and a printing means 24.

The image input means 20 receives digital image data obtained by a scanner, a digital camera, or the like, and outputs the digital image data to the image compression unit 21 as an input image. Here, in addition to directly accepting digital image data,
Receives data in a form that includes control information of the printer device, etc., generates digital image data based on it,
You may make it output as an input image.

The image compression device 21 receives an input image from the image input means 20, performs image compression by the image compression method of the present invention, and outputs the compressed image to the compressed image storage means 22. The image compression device 21 can be realized by the image compression device disclosed in the first embodiment of the present invention.

The compressed image storage means 22 stores the compressed image
1 to receive and store the compressed image. Such a compressed image storage means 22 can be realized by, for example, a semiconductor memory or a hard disk device.

The image decoding means 23 reads the compressed image from the compressed image storage means 22 in response to a request from the printing means 24, executes a decoding process, and outputs the obtained decoded image to the printing means 24. The decoding process extracts, for each compressed partial image, identification information of an image compression algorithm from the compressed partial image, identifies the image compression algorithm used for the compression process based on the identification information, and reverses the compression process by the algorithm. This is performed by performing

For example, if it is determined that the algorithm based on the JPEG method is used, the AC component extracted from the compressed partial image is subjected to Huffman code decoding and run-length compression decoding. A frequency distribution is calculated by performing inverse quantization together, and an inverse DCT is performed on the frequency distribution to decode an image.

If it is determined that the algorithm based on the pallet compression method is used, the pallet number is read for each pixel from the compressed partial image, and the pallet number is stored in the density value stored in the corresponding pallet. By changing, the decoding of the image is performed.

When it is determined that the algorithm using the forced compression method is used, if the number of bits extracted at the time of the forced compression is n, n bits are added to the lower bits for each pixel, and The image is decoded by setting the value of the bit to 0.

The printing means 24 issues a decoding start request to the image decoding means 23 according to the progress of the printing process, receives the decoded image from the image decoding means 23, performs halftoning on the decoded image, and performs halftone data processing. Is generated, and printing is performed based on the halftone data. Conventional techniques such as an error diffusion method, a dither method, and a screen method can be used as a halftoning method.

Next, a third embodiment of the present invention will be described. The third embodiment includes a recording medium on which an image compression program is recorded. This recording medium is CD-RO
M, a magnetic disk, a semiconductor memory, and other recording media, including those distributed through a network.

The image compression program is read from the recording medium into the data processing device, and controls the operation of the data processing device. The data processing device divides the image into partial images of a predetermined size under the control of the image compression program, performs image compression on the divided partial images based on a plurality of image compression algorithms, A good compression result is selected as the compression result of the partial image, and is output together with the identification information of the corresponding image compression algorithm.

That is, the data processing device controls the image compression program in FIG.
0, the same processing as the processing by the compression result storage means 11 and the compression result selection means 12 is executed.

According to the present invention, an image is divided into partial images of a predetermined size, and the divided partial images are subjected to image compression based on a plurality of image compression algorithms. By selecting the result as the compression result of the partial image, image compression is performed using an independent image compression algorithm for each partial image. Even in a mixed image, a high compression ratio can be achieved without deteriorating the image quality.

Further, according to the present invention, the compression result by the palette compression method can be selected as a plurality of image compression algorithms, so that lossless compression can be efficiently performed when the content of the partial image is a character or a figure. It can be carried out.

Further, according to the present invention, as a plurality of image compression algorithms, it is possible to select a compression result by an algorithm that leaves only upper bit information for each pixel, so that the compression rate of such an algorithm is independent of the content of the image. Therefore, the compression ratio can always be set to a certain value or less irrespective of the content of the image.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a first exemplary embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a main part of the first embodiment of the present invention.

FIG. 3 is a schematic diagram for explaining an image compression algorithm according to the JPEG method.

FIG. 4 is a schematic diagram showing a result of compression by a pallet compression method.

FIG. 5 is a diagram illustrating the relationship between the number of types of density values and the compression ratio in the forced compression method.

FIG. 6 is a schematic diagram showing data of a compressed image compressed according to the present invention.

FIG. 7 is a block diagram showing a configuration of a second exemplary embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image compression apparatus 2 Printer apparatus 10 Image division means 11 Compression result storage means 12 Compression result selection means 20 Image input means 21 Image compression means 22 Compressed image storage means 23 Image decoding means 24 Printing means 110 JPEG compression means 111 Pallet compression means 112 Compression means

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5C059 KK25 MA00 MA23 MC04 MC33 MC35 ME02 PP20 PP24 PP29 RC22 RC33 RC34 SS15 SS28 UA39 5C078 AA01 AA04 BA21 BA44 CA02 DA00 DA01 DA02 DB05 9A001 BZ03 EE04 HH27

Claims (9)

[Claims] An image is divided into partial images of a predetermined size,
By performing image compression on the divided partial images based on a plurality of image compression algorithms and selecting a compression result having the best compression rate as the compression result of the partial image, independent image compression is performed for each partial image. An image compression device for performing image compression by an algorithm. 2. An image dividing means for dividing an image into partial images of a predetermined size, and performing image compression on the divided partial images based on a plurality of image compression algorithms, Compression result storage means for storing a plurality of image compression algorithms for each of the plurality of image compression algorithms in association with each other, and referring to the compression result storage means, for each of the divided partial images, Compression result selecting means for selecting a compression result corresponding to the image compression algorithm and outputting it together with identification information of the image compression algorithm. 3. The partial image of a predetermined size is a partial image of 8 × 8 pixels, and the plurality of image compression algorithms are at least JP.
3. The image compression apparatus according to claim 1, further comprising an algorithm based on an EG method, an algorithm based on a pallet compression method, and an algorithm for leaving only upper bit information for each pixel. 4. The method according to claim 1, wherein when the compression ratios of the plurality of image compression algorithms are substantially the same, a compression result corresponding to the image compression algorithm which is a reversible compression is preferentially selected. 4. The image compression apparatus according to 2, 3 or 4. 5. When the compression ratios of the plurality of image compression algorithms are substantially the same, the compression ratios are determined in the order of an algorithm using a pallet compression method, an algorithm using a JPEG method, and an algorithm that leaves only upper bit information for each pixel. 5. The image compression apparatus according to claim 3, wherein the compression result is selected with priority. 6. An image decompression algorithm corresponding to the identified image compression algorithm for each of the divided partial images based on identification information of the image compression algorithm. The image compression apparatus according to any one of claims 1 to 5, further comprising an image decoding unit that performs image decompression based on (i). 7. A printer device comprising the image compression device according to claim 1. 8. An image is divided into partial images of a predetermined size,
By performing image compression on the divided partial images based on a plurality of image compression algorithms and selecting a compression result having the best compression rate as the compression result of the partial image, independent image compression is performed for each partial image. An image compression method characterized by performing image compression by an algorithm. 9. A computer-readable recording medium storing a program for causing a computer to execute the image compression method according to claim 8.