JP2003189108A - Image compression method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a coding method for improving compression rate by applying coding to a pattern consecutively and repetitively appeared. <P>SOLUTION: The image compression method comprises: scanning a binary image constituting an image for each of 8 lines (rows) and encoding parts not including black pixels while skipping them; detecting a unit pattern of a repetitive pattern among the 8 lines being the scanning unit and counting the number of repetitive times when the lines including black pixels have a dither pattern; then encoding the dither pattern by using the number of repetitive times of the unit pattern, a width (the number of columns) of the unit pattern, and pixel data of the unit pattern. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compression method for improving the compression rate of binary bitmap images.

[0002]

2. Description of the Related Art In recent years, information devices such as personal computers have become widespread and used for various purposes. In particular, an image is often processed by a personal computer, and a printer is connected to the personal computer to print an image or transfer image data via a network such as the Internet. However, as the image becomes more complicated, it takes too much time to transfer the image data from the personal computer to the printer or transfer the image data via the network, and it is desirable to compress and transfer the image data. Be done. By compressing the image data, the amount of data to be transmitted is reduced and the transfer time can be saved.

In this case, the image data includes color data, gray scale data, and the like. In the following, a method for compressing a binary bit map image obtained by binarizing these image data will be described.

As a conventional binary bit map image compression method, there are MH (Modified Huffman) and MR (Modified).
READ), MMR (Modified Modified READ), JB
There are IG (Joint Bi-level Image Group) and the like, and they are widely used for storing in files and FAX.

Since these compression methods are complicated in encoding processing, "a method and apparatus for compressing bitmap data" aiming at speeding up of processing speed (Japanese Patent Laid-Open No. 8-51).
No. 545) has been proposed.

In the image compression method of the above-mentioned Japanese Patent Laid-Open No. 8-51545, first, an image is horizontally scanned in a unit of a predetermined number of lines, and blank lines which are continuous between rows are coded as vertical skips. Separate the area as a logical line. Next, the separated logical rows are vertically scanned in units of one column to skip a blank area that does not include black pixels, such as between characters, and separate areas other than the blank area that include black pixels.

Then, an area consisting of only white pixels and an area consisting of only black pixels are coded so as to skip by designating the number of columns respectively, and an area in which columns of the same pixel value are consecutive is determined by the value and the number of repetitions. Encode. In other areas, the pixel data is output as coded data as it is together with the number of columns.

[0008]

The above-described bit map data compression method has a problem that the compression rate of an image in which a pattern having a constant size such as a dither pattern repeatedly appears is reduced.

FIG. 9 is a diagram for explaining a conventional dither pattern coding method. As shown in the figure, the scanning of the image is performed, for example, in units of 8 lines. Since a blank line is a part where there is no image data, information about how many lines of 8 lines are skipped is encoded.

The row containing the dither is recognized as a logical row, and the encoding is performed every 8 lines. Below the figure,
It is a description of a dither pattern encoding method. That is,
Conventionally, a bit pattern is encoded as it is as a dither pattern, and for example, in encoding a logical row consisting of 8 lines × N pixels, a partial area including a dither is encoded as it is. Therefore, the code data is composed of the code word, the pixel width N, and the pixel data (N bytes) of the partial area, which is larger than N bytes.

Therefore, there is a demand for an image compression method capable of effectively compressing an image including a dither pattern. An object of the present invention is to provide a coding method that improves the compression rate by coding patterns that appear repeatedly in succession.

[0012]

An image compression method of the present invention is a method for compressing image data, which comprises an extraction step of extracting a line containing a repetitive pattern from binary image data, Using the unit pattern of the repeating pattern, the repeating pattern detecting step of detecting the number of times of repeating the unit pattern, at least the data of the unit pattern, and the number of times of repeating the unit pattern in the repeating pattern, A coding step for coding a row containing a repeating pattern.

According to the present invention, it is possible to effectively encode image data such as a dither pattern in which unit patterns repeatedly appear, and improve the compression ratio of the image data.

[0014]

1 is a flowchart showing the flow of the entire processing of an encoding method according to an embodiment of the present invention.

First, the input image is scanned in the horizontal direction in units of a predetermined number of lines (step S101), and blank lines which are continuous such as between lines are encoded as vertical skips.
Areas other than blank lines are separated as logical rows (step S102).

Next, the separated logical rows are vertically scanned in units of one column (step S103), and a blank area containing no black pixels such as a space between characters is skipped, and an area containing black pixels other than the blank area is skipped. Are separated (step S104).

Further, based on the result of region separation in step S104, a repeating pattern having the same width and the same pixel value that appears repeatedly and repeatedly is separated (step S10).
5).

Then, steps S102, S104, S
Encoding is performed using each area separated in 105 as an element (S
106). FIG. 2 is a diagram showing the first principle of the embodiment of the present invention.

A code data size is reduced by detecting a pattern which appears repeatedly and repeatedly from a logical row and performing coding by using the length and pixel value of a repeating pattern which is the basis of repetition and the number of times of repetition. To do. Thereby, the code data includes the code word, the number of repetitions K, the pixel width M,
It consists of a basic pattern (M bytes) of a repeating pattern, and the length of the code data of the dither pattern is approximately N / K bytes, which is about 1 / K of the length of the code data of the conventional dither pattern. .

In the process of generating an image, there are cases where the size of the pattern, which is the basic element for repeating the dither pattern used when generating the image, is known. In this case, by limiting the detection of the repetitive pattern to this known size, the detection process can be significantly reduced and detection can be performed at high speed, as compared with the case where detection is performed for all sizes.

FIG. 3 is a diagram showing a second principle according to the embodiment of the present invention. Instead of outputting the detected repetitive pattern as it is as code data, the repetitive pattern itself is compressed by a data compression method such as line length coding, whereby the code data size can be further reduced.

At this time, if there is no compression effect of the repeated pattern, the repeated pattern is output as coded data without being compressed. In the figure, if there is a dither pattern as shown in FIG. 9A, the basic pattern of the repeating pattern having the repeating length of M pixels is repeated K times. Therefore, as shown in (B), a basic pattern of a repetitive pattern consisting of M pixels is taken out and is converted into a run length code. As shown in (C), the run-length coding method examines where the black pixel is in the column in order from the column of the first 8 rows of pixels of the basic pattern. First, since the first four columns in (B) have the same pattern, 1
Information is generated that four pixel patterns for each column are lined up. Next, two columns without black pixels are lined up. In the same manner, a basic pattern of black pixels and information about how many times the pattern is continuous in the column direction are generated for each column.
Generate M ′ bytes of run-length code data. Performing run-length coding provides a much better compression effect than not performing it.

When the run length coding is not performed, the code data is composed of a code word, a run length flag = 0 (indicating that the run length is not coded), a pattern length M, a repeat count K, and pattern data. When the run length encoding is performed, code data having a configuration including a code word, a run length flag = 1 (indicating that the run length encoding is performed), a run length code data length M ′ bytes, the number of repetitions K, and run length code data is obtained. To be

FIG. 4 is a flow chart showing the flow of processing according to the embodiment of the present invention. The flowchart in the figure is
The processing corresponding to the processing of step S105 and step S106 in the principle diagram shown in FIG. 1 and showing the processing of encoding the separated logical row is shown. The blank area is encoded by a conventional method.

The processing flow will be described below. Step S201: Initialization processing such as allocation of a work memory area and initialization of internal parameters necessary before starting the encoding processing is performed. A parameter for counting the number of patterns registered in the pattern table described later is also initialized to zero here.

Step S202: The presence or absence of a pattern having the same size and pixel value that repeatedly appears from the logical row is detected. Step S203: If a repeating pattern is detected based on the detection result of S202, the process proceeds to S205,
If no repetitive pattern is detected, the process proceeds to S20.
Go to 4.

Step S204: Conventional encoding is performed. Step S205: Encode the area of the detected repetitive pattern. Step S206: It is determined whether or not encoding of all logical lines is completed, and if not completed, the process is S202.
And when it is completed, the encoding process is ended.

Hereinafter, each processing of the repeated pattern detection processing S202 and the repeated pattern coding S205 will be described in detail with reference to the drawings. FIG. 5 is a processing flow showing an embodiment of repetitive pattern detection, and corresponds to the processing of S202 of FIG.

The logical rows are scanned in column units to check the presence / absence of a unit pattern of a continuous repeating pattern. If there is a unit pattern of a repeating pattern, the length len and the number of times of repeating together with the unit pattern of the repeating pattern. Get num.

Step S301: num indicating the number of times the unit pattern of the detected repetitive pattern is repeated is set to 0. Also, 1 is set to the variable a indicating the column where the detection of the unit pattern of the repetitive pattern is started, that is, the leading column of the logical row.

Step S302: Iterative pattern for starting detection
MIN for len as turn length Set LEN.
MIN LEN is the unit pattern of the expected repeating pattern.
You can assign an appropriate value from the window. For example, in the image
The width of the unit pattern of the repeating pattern is 4 pixels or more
If you know in advance, LEN = 4
And This will allow you to detect non-existent repetitive patterns.
Output processing can be avoided and processing time can be shortened.

Step S303: A value obtained by adding the unit pattern length len of the repetitive pattern to be detected to a is set to the variable b indicating the reference column number. Step S304: an image in the len column starting with a,
Images in the len column starting from b are compared to determine whether they match. If they do not match, the process proceeds to step S305, and if they match, the process proceeds to step S310.

Step S305: 1 is added to the length len of the unit pattern of the repeating pattern to be detected in the length len column to set the length of the unit pattern of the repeating pattern to be detected to be one column longer.

Step S306: The maximum length MAX detected by the length len of the unit pattern of the repeating pattern. LE
It is confirmed whether or not N has been reached, or len has reached half of the number of remaining columns to which the detection processing has not been applied. If any of the conditions is satisfied, the processing is returned to step S303,
If both conditions are not satisfied, the process proceeds to step S30.
Proceed to 7. Here, the fact that len reaches half of the number of remaining columns (WIDTH-a) to which the detection process is not applied means that the unit pattern of length len starting from the column position a to be subjected to the detection process corresponds to the remaining columns. This means that no more than one can be included in it. Therefore, since it is not a repetitive pattern, the process proceeds to step S307, which is the next pattern detection process.

Step S307: 1 is added to a indicating the repeated pattern detection start sequence. That is, the column for starting the detection of the unit pattern of the repeating pattern is advanced by one column. Step S308: The number of remaining columns to which the detection process is not applied is 2 which is the shortest length of the unit pattern of the repetitive pattern to be detected.
If the number of times remains, the process returns to step S302, and if not, the process proceeds to step S309. In the case of advancing to step S309, the unit pattern to be detected is not included in the remaining column number more than one, so that it is not a repetitive pattern, and therefore the repetitive pattern detection process is terminated. Processing.

Step S309: It is output that there is no repetitive pattern that continuously appears in the logical line, and the process is terminated. By repeating the above processing until YES in step S304, the presence / absence of a repeating pattern that continuously appears in the logical row, and if the repeating pattern exists, the position and the unit pattern of the repeating pattern are detected. If it does, the following processing is further performed.

Step S310: Since it has been detected that two repetitive patterns are continuous by the processing up to this point, 2 is set to the variable num for counting the number of repetitions of the repetitive pattern.

Step S311: b is added to len to indicate the next column for detecting the presence / absence of a repeating pattern, and the number of columns for one pattern is advanced. Step S315: Whether or not the column b, which is the detection processing start position one pattern ahead, is smaller than the column width WIDTH minus the pattern width len, that is,
It is determined whether or not the pattern of length len starting from the column b fits within the remaining number of columns. If b is less than or equal to WIDTH-len, the process proceeds to step S312, and b is WI.
If it is larger than DTH-len, step S314.
Proceed to.

Step S312: Similar to S304, a
Images in the len column starting with b and len starting with b
The images in the columns are compared to determine whether they match. If they do not match, the process proceeds to S314, and if they match, the process proceeds to S313.

Step S313: 1 is added to the counter num indicating the number of repetitions of the detected repetition pattern, and the process is returned to S311. By repeating steps S311 to S313 until no continuous repeating pattern is detected, the number of continuous repeating patterns is detected.

Step S314: The detected start position a of the repetitive pattern, the length len of the repetitive pattern and the number of repetitions num are output as a result, and the process is terminated. Through the above processing, it is determined whether or not a repetitive pattern exists in the logical row. When the repetitive pattern exists, the start position, the length of the unit pattern of the repetitive pattern, and the number of repetitions are obtained.

FIG. 6 is a diagram showing an embodiment of the bit structure of the code word. FIG. 6A is an example in which the detected repetitive pattern is not run-length coded and the repetitive pattern is output as a code as it is, and FIG. 6B is an example in which the repetitive pattern is run-length coded and output. Figure 6
In (a), to indicate that the data is not run-length code data, the run-length coding flag is set to 0, and the unit pattern data of the repetitive pattern having the pixel width designated by the data length M indicating the length of the unit pattern is set. Output. Figure 6
In (b), in order to indicate that the run length encoding is performed, the run length encoding flag is set to 1, and the run length encoded data of the data length M ′ byte indicating the length of the run length encoded data is set. Output. In any case, the number of repetitions K indicating how many times the unit pattern of the repeating pattern appears repeatedly is set in the encoded data.

FIG. 7 is a diagram showing an example of the pattern table. When the dither pattern is encoded, the above-mentioned pattern table registers, for example, a unit pattern appearing in a document of one page together with its length, and when the same unit pattern appears in one page, The registration number of the pattern table is stored in the pattern data or the run length code data portion of the above-mentioned encoded data. The registration number is set sequentially from the unit pattern that appears first while the image is being encoded. The unit pattern that appears first is registration number 1, and the unit pattern that appears second is The pattern is registration number 2.

This pattern table is generated when the image is encoded, but is not included in the encoded data. When decoding the encoded image, the unit patterns are sequentially read from the code data and the pattern table is reproduced on the decoding side. Then, since the unit patterns are embedded in the code data in the order in which they appear at the time of encoding, even when decoding, if the pattern table is created while decoding the code data sequentially, it is possible to The same thing will be obtained. Therefore, when the registration number of the pattern table appears in the code data, the code data is correctly decoded if the unit pattern is acquired by referring to the registration number of the pattern table generated at the time of decoding. be able to.

As described above, by using the pattern table, the compression rate can be increased as compared with the case where the bit data of the unit pattern is directly included in the code data. Furthermore, when used in combination with run length coding, the compression rate of image data can be further increased.

FIG. 8 is a hardware environment diagram required for the information processing apparatus when the method of this embodiment is implemented by a program. R connected to CPU 11 via bus 10
A basic program such as BIOS is stored in the OM 12. The program is executed by the CPU 11 on the bus 10.
It is stored in a storage device 17 such as a hard disk, which is connected to the RAM 13, copied to the RAM 13, and executed by the CPU 11. The reading device 18 reads the program from a portable recording medium 19 such as a flexible disk, a CD-ROM, a DVD or the like, and enables the information processing device 21 to be installed. Alternatively, by directly reading the program from the portable recording medium 19, the CPU
11 may be executed.

The input / output device 20 includes a keyboard, a mouse,
It is composed of a template, a display, etc., and transmits a command from the user of the information processing device 21 to the CPU 11 or C
It is for presenting the calculation result of the PU 11 to the user.

The communication interface 14 connects the information processing device 21 to the information provider 16 via the network 15. The information provider 16 has the program, downloads the program to the information processing device 21 via the network 15, installs the program, and then the CPU
11 may execute. Alternatively, while the information provider 16 and the information processing device 21 are connected via the network 15,
The program may be executed under a network environment.

(Supplementary Note 1) A method of compressing image data, which comprises a step of extracting a line containing a repeating pattern from binary image data, a unit pattern of the repeating pattern, and a unit pattern of the repeating pattern. Of a repetition pattern detecting step of detecting the number of repetitions of the unit pattern, at least data of the unit pattern, and the number of repetitions of the unit pattern in the repetition pattern And an image compression method.

(Supplementary Note 2) The image compression method according to Supplementary Note 1, wherein the pixel width of the unit pattern is one or more predetermined fixed values. (Supplementary note 3) The image compression method according to Supplementary note 1, wherein in the encoding step, the unit pattern is further compressed, and encoded by the compressed unit pattern and the number of repetitions of the unit pattern.

(Supplementary Note 4) The image compression method according to Supplementary Note 1, wherein the encoding step uses a run length encoding method. (Supplementary note 5) In the encoding step, a table for registering the unit patterns that have appeared is generated, and when a unit pattern that is the same as the unit pattern appears twice or more in the process of encoding, the unit pattern is replaced with the unit pattern. 2. The image compression method according to appendix 1, wherein the image data is encoded using the registration number of the unit pattern in the table.

(Supplementary Note 6) A method of compressing image data, comprising a step of extracting a line containing a repeating pattern from binary image data, a unit pattern of the repeating pattern, and a unit pattern of the repeating pattern. Of a repetition pattern detecting step of detecting the number of repetitions of the unit pattern, at least data of the unit pattern, and the number of repetitions of the unit pattern in the repetition pattern A program for causing an information processing apparatus to realize an image compression method, characterized by comprising:

(Supplementary Note 7) A method for compressing image data, comprising a step of extracting a line containing a repeating pattern from binary image data, a unit pattern of the repeating pattern, and a unit pattern of the repeating pattern. Of a repetition pattern detecting step of detecting the number of repetitions of the unit pattern, at least data of the unit pattern, and the number of repetitions of the unit pattern in the repetition pattern An information processing device-readable recording medium storing a program for causing an information processing device to realize an image compression method, the method comprising:

(Supplementary Note 8) A device for compressing image data, comprising: an extracting means for extracting a line in the binary image data that includes a repeating pattern, a unit pattern of the repeating pattern, and the unit pattern. Code for encoding a row containing a repetitive pattern in image data using at least the data of the unit pattern and the number of repetitions of the unit pattern in the repetitive pattern. An image compression apparatus comprising: an image compression unit.

[0055]

As described above, according to the image compression method of the present invention, the basic pattern which is the basis of the repetition is detected and encoded from the image in which the same pattern continuously appears repeatedly such as the dither pattern. By doing so, the compression rate of images including dither such as photographs is improved. Therefore, it greatly contributes to the reduction of the storage capacity for storing the data and the reduction of the data transfer time.

[Brief description of drawings]

FIG. 1 is a flowchart showing the overall flow of processing of an encoding method according to an embodiment of the present invention.

FIG. 2 is a diagram showing a first principle of the embodiment of the present invention.

FIG. 3 illustrates a second principle according to an embodiment of the present invention.

FIG. 4 is a flowchart showing a flow of processing according to the embodiment of the present invention.

FIG. 5 is a process flow showing an embodiment of repetitive pattern detection.

FIG. 6 is a diagram showing an embodiment of a bit configuration of a codeword.

FIG. 7 is a diagram showing an example of a pattern table.

FIG. 8 is a hardware environment diagram required for the information processing device when the method of the present embodiment is implemented by a program.

FIG. 9 is a diagram illustrating a conventional dither pattern encoding method.

[Explanation of symbols]

10 bus 11 CPU 12 ROM 13 RAM 14 Communication interface 15 network 16 Information provider 17 Storage 18 Reader 19 Portable recording medium 20 I / O device 21 Information processing equipment

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 5C078 AA01 AA06 BA35 CA01 DA01                       DA05                 5J064 AA01 BA08 BB05 BC01 BC05                       BC14 BC29 BD02

Claims (5)

[Claims] 1. A method of compressing image data, comprising: an extracting step of extracting a row containing a repeating pattern from binary image data; a unit pattern of the repeating pattern; and repeating the unit pattern. Using a repeating pattern detection step of detecting the number of times, at least the data of the unit pattern, and the number of times the unit pattern is repeated in the repeating pattern,
An encoding step of encoding a row containing a repetitive pattern in the image data, the image compression method. 2. The image compression method according to claim 1, wherein the pixel width of the unit pattern is a predetermined fixed value of 1 or more. 3. The image compression according to claim 1, wherein, in the encoding step, the unit pattern is further data-compressed and is encoded by the compressed unit pattern and the number of repetitions of the unit pattern. Method. 4. The image compression method according to claim 1, wherein the encoding step uses a run-length encoding method. 5. In the encoding step, a table for registering the appeared unit patterns is generated, and when the same unit pattern appears twice or more in the encoding process, the unit pattern of the unit pattern is generated. The image compression method according to claim 1, wherein the image data is encoded by using the registration number of the unit pattern in the table instead.