JP2000236448A - Data compressor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the compressibility of multi-valued image data by outputting a binarize-expressed gradation pattern corresponding to a pixel value and selecting and compressing a gradation pattern corresponding to the pixel value so as to increase regularity seen from an arraying direction. SOLUTION: Multi-valued image data are accepted from an input device 1 to be compress-processed and outputted to an outputting device 2. Then, a pixel scanning part 3 fetches each pixel value included in a multi-valued image. A pattern conversion part 4 accepts the pixel value outputted by this part 3 and outputs the binarize-expressed gradation pattern corresponding to the pixel value. A data compressing part 5 arrays the gradation patterns outputted by the part 4 in the order of outputting and executes data compression by using regularity seen from the arraying pattern. In this case, the part 4 outputs the gradation patterns corresponding to each pixel value which is obtained by selecting dot array so as to increase regularity seen from the arraying direction.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a data compression apparatus for receiving multi-valued image data, compressing the data, and outputting the data.

[0002]

2. Description of the Related Art When image data created by a host device is transferred to an output device or stored in a storage device, data compression is performed to reduce the amount of data. Methods such as run-length encoding, Huffman encoding, and MMR encoding are known for compressing binary image data composed of only white or black pixels.

In the run-length encoding method, a block (run) of continuous white pixels or black pixels is detected for each raster line of image data to be compressed. Then, an arbitrary code is assigned to each length of the run, and the arrangement of all the pixels in the raster line is described by a sequence of these codes. A short bit code is assigned to a run length that frequently appears, and a long bit code is assigned to a run length that appears less frequently. Thus, the data amount of one raster line can be compressed. Such techniques are widely used in, for example, facsimile machines.

On the other hand, compression of multi-valued image data is performed by JPE
The G method and the Ziv-Lempel method are known. The former is an irreversible compression method that is suitable for compression of moving images and still images, but cannot completely reproduce data before compression from data after compression. The latter is a method suitable for compressing character code data of a document or the like.

[0005]

However, the above-mentioned prior art has the following problems to be solved. For example, when print data is transmitted from a higher-level device to a printer, a method of temporarily compressing the print data, restoring it after receiving it by the printer, and the like is adopted. In this case, reversible data compression is required. Also,
Print data transferred to a printer generally includes a large number of character images. The multi-valued image data compression method for processing a continuous tone image such as a photograph has a problem that image quality degradation is conspicuous for such a character image. In addition, complicated arithmetic processing is required, and the bit length of an applicable image is limited.

Therefore, one multi-valued image is decomposed into a plurality of binary bit plane images. That is, a bit plane image in which only the most significant bit of each pixel value of the multi-valued image is collected,
A bit plane image corresponding to the number of bits of the pixel value is prepared in a manner such as a bit plane image in which only the second bit is collected. Then, each bit plane image is compressed using a binary image data compression method.

However, there is a problem that a sufficient compression ratio cannot be obtained because each bit of the lower-order bit plane image has large variations. In addition, a large amount of memory is required for decomposing into a bit plane image, and the maximum number of gradations, that is, the number of bits for representing pixel values is also limited. Therefore, when compressing multi-valued image data, the compression ratio is higher,
It is desired to establish a compression method that is not restricted by the maximum number of gradations.

[0008]

The present invention employs the following structure to solve the above problems. <Structure 1> A pixel scanning unit for extracting each pixel value included in the multi-valued image data, and accepting a pixel value output from the pixel scanning unit, and outputting a binary gradation pattern corresponding to the pixel value And a data compression unit that arranges the gradation patterns output by the pattern conversion unit in the output order, and compresses the data using regularity seen in the arrangement direction. And outputting a gradation pattern corresponding to each pixel value, wherein the dot arrangement is selected so that the regularity seen in the arrangement direction becomes large.

<Structure 2> In the data compression device according to Structure 1, the pattern conversion unit may increase the probability that pixels of the same type are arranged in the arrangement direction when the gradation patterns are arranged in the output order. A data compression device for outputting a gradation pattern corresponding to each pixel value in which a dot array is selected.

<Structure 3> In the data compression apparatus according to Structure 1, the pattern conversion section may increase the probability that dots of a certain combination are arranged in the arrangement direction when the gradation patterns are arranged in the output order. A data compression apparatus for selecting a gradation pattern corresponding to each pixel value in which a dot array is selected.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below using specific examples. FIG. 1 is a block diagram showing a data compression device according to the present invention. This device is configured to receive multivalued image data from an input device 1, perform compression processing, and output the compressed data to an output device 2. The input device 1 is, for example, an application or a higher-level device that generates multi-valued image data, and the output device is, for example, a storage device or a printer. This device includes a pixel scanning unit 3, a pattern conversion unit 4, a binary data compression unit 5, and a pattern conversion table 6. The pattern conversion table 6 stores a gradation pattern 7 as shown in the upper part of the figure.

In this example, for example, each pixel value of the multi-valued image data is represented by eight types from 0 to 7. The binary gradation pattern corresponding to the pixel value is stored in the pattern conversion table 6. That is, the pattern conversion table 6 stores eight types of gradation patterns 7. The gradation pattern 7 in the figure is constituted by a 3 × 3 matrix, and the numerical value written in each cell indicates the order of the pixels that are switched from white pixels to black pixels each time the pixel value increases by “1”. ing.

That is, for example, when the pixel value is "1",
Only the portion of the upper right corner of the gradation pattern 7 in which the numeral "1" is entered becomes a black dot, and the other portions become white dots. When the pixel value is “2”, the portion where the numerical value of the upper right corner and the upper center is written as 1 and 2 in the row of (1) becomes a black dot, and the other portions become white dots. Input device 1
, And the arrangement direction of the above gradation patterns 7 arranged in order in the image data is indicated by an arrow A in the figure.

In this case, the gradation pattern 7 as described above is used.
Is adopted, when the pixel value is “3” or more, all the dots in the row shown in FIG. If the pixel value is less than "7", all the dots in the row shown in (3) are white dots. Therefore, data in which a large number of such tone patterns 7 are arranged in the direction of arrow A has a high probability of occurrence of black runs and white runs. That is, there is a high probability that black dots and white dots are continuously arranged in the arrangement direction. Therefore, if the data is compressed by a conventional run length method or the like, the data can be compressed at a high compression rate.

FIG. 2 shows a data compression method and a modified explanatory diagram of a gradation pattern. FIG. 2A shows an example of the binarized image data 8 arranged in one horizontal line. As shown in the figure, for example, data such as a white run having a length of “4”, a black run having a length of “3”, a black run having a length of “5”, and a white run having a length of “1” are included. In this case, a fixed code 5 is applied to each run length as shown in FIG.

In the example shown in FIG. 2B, the run length is "1".
Is "0000", and when the run length is 16, the code word is "1111". Therefore, the longer the run length, the higher the compression ratio. For this reason, when the gradation patterns 7 as described above are arranged long in the direction of the arrow A shown in FIG. 1, compression at a high compression ratio becomes possible. Note that the above-described method of selecting a gradation pattern so as to increase the probability that pixels are arranged can be modified as shown in FIGS. 2C to 2F, for example.

(C) is the same as the example shown in FIG.
(D) shows an example in which the right and left sides of (c) are reversed.
(E) is an example in which (c) is upside down, and (f) is an example in which (e) is upside down. In any case, a high compression ratio can be obtained when the gradation patterns 7 are arranged in the direction of arrow A shown in FIG.

FIG. 3 is a diagram for explaining the regularity in the arrangement direction. The effect of improving the compression ratio when the above gradation patterns are arranged will be described with reference to FIG. The original image data 11 shown in the figure has a structure in which each pixel value of the multi-valued image data is indicated by a numeral. As in the examples so far, the pixel value of one pixel is selected in the range of 0 to 7. When the pixel value of each pixel is converted into a binary tone pattern as described above, the tone patterns are arranged as shown on the right side or the lower side of the figure. The right side shows image data of an image to be printed on portrait format paper.

In this example, each gradation pattern is sequentially arranged in the X direction. The gradation pattern composed of 3 × 3 dots is arranged as shown in the figure by a combination of corresponding black dots and white dots. As can be seen from this example, for example, when the pixel values are arranged as 5, 6, and 5, there are many portions where black dots continue in the X direction and portions where white dots continue in the X direction. Thus, when data compression is performed in the X direction, data can be compressed at a high compression rate.

Conventionally, the gradation pattern is set so that the positions of the black dots are dispersed as much as possible. Therefore, data after binarization using a gradation pattern tends to have a very low compression ratio. In the present invention, as shown in this figure, the dot arrangement is selected so that the probability that pixels of the same type are arranged in the arrangement direction of the gradation pattern is increased, and the compression ratio is increased.

The paper 13 shown on the lower side of the figure is a landscape type paper. Even in such a case, for example, when the gradation patterns are arranged and output in the Y direction, this Y
The dot arrangement is selected so that the probability that pixels having the same value are aligned in the direction is high. Thereby, a similar effect can be obtained.

In the above example, the example in which the probability that pixels having the same value are arranged is increased. However, for example, the probability that dots of a certain combination are arranged in the arrangement direction may be increased. . That is, when a certain pattern composed of a combination of white dots and black dots is repeated many times as viewed in the arrangement direction, a method of coding and compressing the repetition state of such a pattern may be considered. For example, "0
A fixed code is determined according to the number of times the pattern of "11" is repeated. In this way, data can be compressed.

The specific operation of each part of the apparatus shown in FIG. 1 will be described below. FIG. 4 is a diagram illustrating the operation of the pixel scanning unit 3. The pixel scanning unit 3 takes out the signals input from the input device 1 in an array according to the request of the output device,
It has a function of outputting this to the pattern conversion unit 4. For example, when the output device is an electrophotographic printer, FIG.
As shown in (a), a data array that scans one line at a time in the width direction of the sheet is required. That is, data is taken out and output one line at a time so that main scanning is performed in the width direction of the paper and sub scanning is performed in the height direction.

On the other hand, in the case of a so-called serial printer such as an ink jet printer or a dot impact printer, printing is performed by moving a head having a nozzle group arranged in the sub scanning direction in the main scanning direction. Therefore, the order for supplying data differs. That is, FIG.
As shown in (b), a scan is performed in which a pixel data group corresponding to the height of the head is taken out, and this operation is sequentially repeated in the width direction. After extracting the band-like pixel data group corresponding to the height of the head, the head is shifted in the height direction by exactly the height of the head, and a similar scan is performed.

The pixel scanning unit 3 performs, for example, such an operation to extract each pixel value included in the multi-valued image data and transfer it to the pattern conversion unit 4. Pattern converter 4
Outputs a gradation pattern as shown in FIG. 1 corresponding to each pixel value. The binary data compression unit 5 regards the image data output from the pattern conversion unit 4 as a set of binary data, and compresses the data in the arrangement direction of the binary data.

FIG. 5 shows the pattern converter 4 shown in FIG.
2 shows an operation flowchart. First, in step S1, the coordinate values of pixels to be read from the pixel scanning unit 3 are initialized. Then, in the next step S2, 1
The processing for taking in the pixels for the row is performed. In the next step S3, the processing of step S2 is repeated until pixel reading for one page is completed line by line.

FIG. 6A shows a subroutine flowchart of step S2. First, in step S11 in FIG. 6, a pixel specified by X and Y coordinate values is scanned, and in step S12, the pixel value is stored in a scan memory. The scan memory is a memory in which pixel values are arranged and stored in the order of reading, as shown in FIG. In the figure, a memory having a horizontal X and a vertical Y shape is illustrated in consideration of a print image. However, the memory only needs to be able to serially and continuously store all input data.

In step S13 in FIG. 6, X is incremented, and in step S14, control is repeatedly performed until scanning of one row is completed. When the processing for one row is completed, X is initialized to "0" and Y is incremented in step S15. Thus, the process proceeds to step S3 in FIG. By the above processing,
When the pixel values for one page are stored in the scan memory as shown in FIG. 6B, the maximum gradation number, image data width and image data height are set in step S4 in FIG. Then, in step S5, the values of X and Y are initialized.

In the next step S6, a process of converting the pixels into gradation patterns for each row is executed. Step S7
Is a part for performing control to repeat step S6 until the gradation pattern conversion processing of this pixel is performed for one page.

FIG. 7A is a flowchart of a subroutine for patterning one row of pixels. FIG.
In step S16, the pixel values stored in X and Y in the scan memory are read. Then, in step S17, the read multi-valued image data is converted into a gradation pattern as described above and stored in the pattern memory.

FIG. 7B illustrates the structure of the pattern memory. As shown in FIG.
Storage area. Also in this case, the memory may be any memory as long as it can take out data serially in the order of input, and it is not necessary to pattern the data vertically and horizontally. The contents of each gradation pattern are as described above.

In step S18 in FIG. 7, it is determined whether or not one row has been completed, and conversion to a gradation pattern is executed for all pixel values. When the processing for one row is completed, the process proceeds to step S20, where X is initialized to “0” and Y is incremented. FIG. 5 shows such processing.
Is repeated for one page under the control of step S7. 1
When the processing for the page ends, the operation of the pattern conversion unit ends.

In step S8, X and Y are initialized to "0", and in step S10, compression processing is started. In this case, the compression processing is advanced line by line from the top of the pattern memory shown in FIG. Then, in step S10, the repetition processing for one page is controlled. FIG. 7C shows a subroutine of this compression processing.

In step S21 of FIG. 7, the Y-th row of the pattern memory is called, compressed and output by the run-length method. In step S22, Y is incremented. In this way, the compression process from Y to “0” to the last line is executed.

By performing the above control, the compression of the target binarized image data is completed. Therefore, for example, in the case of the electrophotographic system shown in FIG. 4, data is compressed for each line in the width direction, and in the case of an impact type printer, data is compressed by the length of the head.
Is output to the output device 2 shown in FIG. On the output device 2 side, the received multi-valued image data can be restored by decoding the received data using a corresponding algorithm and then performing an inverse conversion using the same gradation pattern 7 as in the pattern conversion unit.

[0036]

As described above, according to the present invention, each pixel value included in the multi-valued image data is extracted in the output order of the apparatus, and a binary gradation pattern corresponding to the pixel value is output. Since the gradation pattern corresponding to the pixel value is selected and compressed so that the regularity seen in the arrangement direction is increased, the multi-valued image data can be efficiently compressed at a high compression ratio. Become. In addition, this makes it possible to compress and output data with high efficiency simply by setting the contents of the pattern conversion table 6 without being affected by the maximum number of gradations of multi-valued image data.

[Brief description of the drawings]

FIG. 1 is a block diagram of a data compression device according to the present invention.

FIG. 2 is a diagram illustrating a data compression method and a modification of a gradation pattern.

FIG. 3 is an explanatory diagram of regularity in an arrangement direction.

FIG. 4 is a diagram illustrating the operation of a pixel scanning unit.

FIG. 5 is an operation flowchart of a pattern conversion unit and below.

FIG. 6 is a subroutine flowchart (part 1).

FIG. 7 is a subroutine flowchart (part 2).

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input device 2 Output device 3 Pixel scan part 4 Pattern conversion part 5 Binary data compression part 6 Pattern conversion table 7 Gradation pattern

Claims (3)

[Claims] 1. A pixel scanning unit for extracting each pixel value included in multi-valued image data, and a pixel value output from the pixel scanning unit is received, and a binary gradation pattern corresponding to the pixel value is formed. A pattern conversion unit for outputting, and a data compression unit for arranging the gradation patterns output from the pattern conversion unit in an output order and compressing the data using regularity seen in the arrangement direction, Is a data compression apparatus which outputs a gradation pattern corresponding to each pixel value, wherein a dot arrangement is selected so that regularity as viewed in the arrangement direction is large. 2. The data compression apparatus according to claim 1, wherein the pattern conversion unit is configured to perform dot formation such that when the gradation patterns are arranged in the output order, the probability that pixels of the same type are arranged in the arrangement direction is high. A data compression apparatus for outputting a gradation pattern corresponding to each pixel value in which an arrangement is selected. 3. The data compression device according to claim 1, wherein the pattern conversion unit increases a probability that dots of a certain combination are arranged in the arrangement direction when the gradation patterns are arranged in an output order. A data compression device, wherein a dot pattern is selected, and a gradation pattern corresponding to each pixel value is selected.