JPS61166273A - Multi-value image data compression system - Google Patents

Abstract

PURPOSE:To increase the substantial storage amount and to decrease the transmission time by checking a binary image of each plane so as to discriminate whether or not the data quantity is compressed, outputting a digital image data as to a plane no data is compressed and outputting the data while compressing as to a plane only whose data is compressed. CONSTITUTION:A data of each bit plane is inputted to an input terminal 1 at each scanning line. The run length of an inputted line data is measured depending whether the color is white or black by each color run length counter 2 or 3. The result of each run length is compared with a value n1 or n2 by a comparator 4 or 5 so as to discriminate whether the white run length is larger than the n1 or the black run length is larger than the n2. When it is discriminated to be larger, the result is counted by an accumulation counter 7 via an OR circuit 6. When the input of the final line of a bit plane is finished, the numeral value of the accumulated counter 7 is compared with n3, and when the value is the n3 or over, the data is outputted as it is without being compressed at recording or transmission.

Description

[Detailed Description of the Invention] Industrial Application Field The present invention is applicable to storing digital image data in a storage medium,
This invention relates to a multivalued image data compression method for transmission.

Digital image data including conventional technical gradation is, for example, 8 dots/
A4 size 1 with +nm resolution and 8-bit gradation
The amount of image data for a page is equivalent to 4MB.

To transmit such data, the transmission speed is 9600
If it is set to bps, it takes about 1 hour to transmit one image. Furthermore, even if it is recorded on a recording medium, it requires 4 MB, and even an optical disk memory, which is currently attracting attention as a large-capacity storage medium, is about IGB, so it can only store about 250 sheets. Therefore, if large amounts of data can be compressed, the transmission time will be shortened, and the number of images that can be stored will be increased.
This is a necessary technology. In facsimiles that currently handle black and white binary values, many of the documents handled are black characters written on white paper, so this property is used to provide codes and compress data. That is, image data is broken down into scanning lines, the number of consecutive dots of a white signal (generally referred to as a run) and the number of consecutive dots of a black signal in each scanning line are counted, and a code is given based on the run length. At this time, because of the accuracy of black characters on a white background, a code with a short length is assigned to a short black run length. Runs with a statistically high probability of occurrence are assigned short codes, and runs with a low probability of occurrence are assigned long codes to reduce the total amount of data required.

This encoding is called Langrenge coding, and in fact, in facsimiles and the like, the MH code, which is a slightly modified version of this encoding method, is used, which is recommended internationally by CCITT. The encoding performed for each scanning line as described above is considered to be one-dimensional data compression.

Also, letters and figures have a correlation power of 1 with the next line, so
Two-dimensional data compression is also performed in which a code is assigned to changes from line to line (MR code).

As mentioned above, international recommendations have been made regarding black and white binary image data (b), and it is becoming established as an international standard. However, compression of image data having intermediate gradations is still at the research stage, and a unified compression method has yet to be determined, but there are many methods that have been proposed. For example, "Image Processing for If, rFAX, OA" (by Takahiko Fukinuki, published by Japan-FIJ Kogyo Shinbunsha, 1983), page 134, is described here.

In that, each bit plane in the gradation direction is black and white 2
It is possible to apply a black and white binary data compression method to the image as a value image. This is called bit-plane encoding and is detailed in the above-mentioned literature. The problem with this encoding is that the higher bit planes are less effective in compression, but the lower the bits, the worse the effect becomes, and conversely, the amount of data may increase. This situation will be explained with reference to FIG. 3 and subsequent figures. Figure 3 is a diagram showing a gradation image, which is shaped like a pyramid and shows how the density becomes black towards the apex, and the numbers written inside represent the density, and the density level is 1'0''.
It is expressed in 8 gradations from 7" to 7" (2N, N is a natural number, here N = 3).The output is a density level of 2
It is a diagram expressed in base numbers. FIG. 5 is a diagram showing line data, in which the vertical axis represents the density level and the horizontal direction represents the line direction. The table at the bottom shows data in bit plane O, bit plane 1, and bit plane 2.

The line data in FIG. 5 represents one line of data that cuts the vertex in FIG. Figure 6 (A) (B) (C)
is a diagram representing bit planes, each with 11 OI
A binary value of t"1" is displayed. (A) is plane 2,
(B) represents plane 1, and (C) represents plane O. Each figure represents 111" as black,
In the upper bit plane 2, black exists as a cluster in the center, so binary data compression has a compression effect. However, since black and white areas are scattered in the lower bit plane O, there is a possibility that the amount of data will increase. As described above, since the compression effect of bit-plane encoding differs from plane to plane, there is no guarantee that the total amount of data will be completely compressed compared to the original amount of data.

Problems to be Solved by the Invention In such conventional bit-plane encoding, a binary data compression method is uniformly applied to each bit-plane.
There is no guarantee of its compression effectiveness.

An object of the present invention is to provide a multi-valued image data compression method that can record or transmit data with the smallest amount of data by automatically determining whether compression is effective by compressing each bit plane. .

Means for Solving the Problems The multivalued image data compression method of the present invention is based on 2 to the Nth power (however,
When recording or transmitting digital image data expressed by N bit planes with gradations (N is a natural number), the binary image of each plane is examined to determine whether the amount of data can be compressed, and the data A plane in which the amount of data is not compressed outputs the digital image data as is, and only a plane in which the amount of data is compressed compresses and outputs the data.

Effect: With this configuration, data that is judged not to be reduced even if the data is compressed is output without data compression, which reduces data storage and transmission time compared to conventional methods that compress everything. This allows the amount of data to be reduced.

Example Hereinafter, one embodiment of the present invention will be described based on the drawings.

FIG. 1 shows a first embodiment. (1) is an input terminal to which line data is input, (2) is a counter that counts the white run length, and (3) is a counter that counts the black run length. (4)
(5) (8) is a comparator, (6) is an OR circuit, (7)
is a total counter, and (12) is an output terminal for the determination result. Comparators (4), (5), and (8) are each given numerical values to be compared.

Next, the configuration of FIG. 1 will be explained in detail based on the operation. The data of each bit plane is input to the input terminal (1
) is input. The run length of the input line data is measured by a run length counter (2) or (3) for each color, depending on whether the color is white or black. The run length is determined when a color opposite to the color currently being counted appears, and at the same time, the counter for the opposite color starts operating to begin measuring the run length of the next opposite color. Each run length measurement result is compared with the value of nl or n2 in comparators (4) and (5), respectively, and it is determined whether the white run is larger than n or the black run is larger than n2. It is carried out. When it is determined that it is large, the OR circuit (6)
is counted by the cumulative total counter (7).

The OR circuit (6) is inserted to operate the cumulative counter (7) when both the white run and the black run are larger than the reference n or n2.

This operation is performed for all lines of the bit plane, and the value of the cumulative counter (7) at the time when the input of the last line of the bit plane is completed is n
This is the sum of the number of white runs greater than or equal to 1 and the number of black runs greater than or equal to n2. This is compared with n3, and when it is greater than or equal to n3, it is determined that this pill-plane should not be compressed and is not stored or transmitted. In some cases, the data is output as is without data compression.

In this way, in the embodiment of FIG. 1, the number of white runs greater than or equal to n1,
If the total number of black runs equal to or greater than n2 is equal to or greater than a certain value n3, that bit plane can be compressed. The specific numerical values n□, n2, and n3 must be varied depending on the type of image being handled, but they can also be determined statistically.

FIG. 2 is a diagram showing a second embodiment. Components in FIG. 2 that have the same functions as those in FIG. 1 are given the same reference numerals and their explanations will be omitted. (13) is a line memory, (14) is a horizontal mode detection circuit, (15) is a counter, and (16) is a comparator, which are the same as those shown by (4) or (5) in FIG. Now let me explain the operation. In addition to the MW code using the Langrenge code, a modified read (MR) code is a two-dimensional compression method that is internationally recommended by the CCITT. To put this simply,
Each scan line has a correlation with the previous line, and parts that change the same as all lines are given a short code meaning the same as the previous line, and if it is within ±3 dots of the run length, it is Only the difference is encoded, and if it is more than that, it is encoded with MH code. That is, if the pattern is the same as the previous line, it is encoded with a short code, and if it is different, it is encoded with an MH code. However, it is mandatory to send one line per fixed line using the MH code, including the possibility of transmission or equipment malfunctions. This is because in the -dimensional case, the error ends in one line, but in the two-dimensional case, it spreads to one line after another. Now, in the case of the present invention, each bit plane of halftone image data is regarded as a black and white binary image, and when determining whether to perform binary encoding on each plane, A dimensional data compression method is applied, but in order to automatically determine whether the amount of data after encoding increases or not regarding the data compression,
The total number of occurrences of the horizontal mode in which the code length of the MR code is most likely to be the longest is measured, and the suitability of compression for that plane is determined based on the cumulative value. Similar to FIG. 1, in FIG. 2, a line-by-line signal obtained by decomposing each plane, which is regarded as a black-and-white binary image, into each scanning line is input to an input terminal (1). The operations of the white run counter (2) and black run counter (3) are the same as in the first embodiment, but in the secondary MR code, the forced MH coding line, the first row of each plane, and the MH of every fixed line are Measurement is performed only on the encoded line. MR encoding requires comparison with the data of the previous line, but the input data is input to the line memory (13) at the same time as it is input.

The input data is compared between the input data of the next encoded line and the data of the previous line in the line memory (13). This is done within the horizontal mode detection circuit, which provides an output signal each time a horizontal mode is detected. The output is counted by a counter (5).

When one line of data has been inputted, the result of the counter (15) at that time is the number of times the horizontal mode has occurred for that line, and that value is set by the comparator (16) as the set value n4.
It is compared with n4 and is output only when it is larger than n4. This output is input to the OR circuit (6) and output to the prime side counter (7) in the same way as the comparators (4) and (5). The cumulative counter (7) sequentially counts each output of the OR circuit (6), and when a plane of input data is input, the numerical value of the counter (7) is for that plane. The sum of the number of occurrences of a white run length of n1 or more, the number of occurrences of a black run length of n2 or more, and the number of occurrences of a horizontal mode of n4 or more is obtained, and this is input to a comparator (8). The comparator (8) compares it with the numerical value of n3, and when it is greater than or equal to n3, outputs a determination result (12) for determining the suitability of data compression in this plane. This operation is repeated for each plane, and by determining whether or not to perform monochrome binary data compression for each bit plane, data compression of halftone image data can be performed. As a result, the transmission time can be shortened during transmission, and the number of stored screens can be increased when storing.

Effects of the Invention As described above, the multilevel image data compression method of the present invention examines the binary image of each plane to determine whether or not the amount of data can be compressed. The image data is output as is, and only the plane where the amount of data is compressed compresses and outputs the above data.
Compared to the case where all the data is compressed and recorded or transmitted as in the embodiment, it is possible to record or transmit a smaller amount of data, thereby increasing the actual storage amount and shortening the transmission time.

[Brief explanation of drawings]

FIG. 1 is a block diagram of main parts of a specific embodiment of the present invention,
Fig. 2 is a block diagram of main parts of another embodiment, Fig. 3 is an explanatory diagram of a halftone image, Fig. 4 is a binary representation diagram of Fig. 3, and Fig. 5 is an explanatory diagram of density level and bit plane. , FIG. 6 is an explanatory diagram of a bit plane. (1)...Input terminal, (2)...White line counter, (3)...Black line counter, (4) (5) (
8) (16)...Comparator, (6)...OR circuit,
(7)... Cumulative counter, (12)... Judgment result output terminal, (13)... Line memory, (14)...
...Horizontal mode detection circuit, (15)...Counter agent Yoshihiro Morimoto 4th cause 5th cause 6th ward

Claims (1)

[Claims] When recording or transmitting digital image data expressed by N bit planes having gradations of 1 and 2 to the Nth power (where N is a natural number), the binary image of each plane is examined. Multi-valued image data is determined whether the data amount is compressed or not, and the plane where the data amount is not compressed is outputted as it is, and only the plane where the data amount is compressed is compressed and outputted. Compression method. 2. When using a one-dimensional run-length code as the data compression method, determine whether the amount of data can be compressed or not.
Measuring the cumulative frequency of occurrence of white or black run lengths in a binary image for each bit plane, comparing this cumulative value with a reference value, and determining that compression is not possible when the cumulative value is greater than or equal to the reference value. A multivalued image data compression method according to claim 1, characterized in that: 3. When using the two-dimensional modified read method as the data compression method, the judgment as to whether or not the amount of data is compressed is made by calculating the cumulative frequency of occurrence of white or black run lengths in the binary image for each bit plane. 2. The multivalued image data compression method according to claim 1, wherein each scanning line is compared with the previous line to determine whether or not compression is to be performed by measuring the cumulative total of different run lengths.