JPH05308505A - Method and device for reducing quantity of image information - Google Patents

Abstract

PURPOSE:To compress redundancy by dividing a source image into a blank part A where all the picture elements are ''0,'' and constructural blocks B where one or more picture elements not being '0' are contained and a main scanning direction and a sub scanning direction are separated just for the prescribed number of picture elements or more from each other, and replacing the information of the part A with the positions and heights of the blocks B. CONSTITUTION:It is discriminated whether all the picture elements of respective lines in the source image are '0' or one or more picture elements not being '0' are contained and when one or more picture elements not being '0' are contained, the respective lines are classified into the blank block where all the picture elements are '0', and image blocks which contain one or more picture elements not being '0' and have the integral multiple length of a prescribed value in the main scanning direction. Next, the leading positions and width of the respective image blocks are successively and repeatedly measured from the heads of the lines concerning all the lines and by using this result, the lines are divided into the blank part A where the picture elements are '0', and the structural blocks B where one or more picture elements not being '0' are contained. Then, the information of the A is replaced with the positions, width and heights of source images in blocks B1-B3. Thus, the quantity of redundant image information is reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for reducing the information amount of an image, and more particularly to a method and apparatus for reducing the information amount of an image for reducing the redundant information amount of the image.

[0002]

2. Description of the Related Art Conventionally, image data has been subjected to processing such as image compression when stored in a storage device such as a memory. There are various compression methods, but most of them encode the entire image.

By the way, the normal image data as shown in FIG. 1 can be divided into a portion B of characters, photographs, figures and the like and a blank portion A as shown in FIG. 2, and the blank portion A is a redundant portion. .. 1 and 2, a thin line represents a portion having non-zero pixels, and a thick line represents a boundary between the margin portion A and the portion B such as a character, a photograph or a figure. If the amount of information in the blank area A is reduced in advance, the amount of information in the image to be processed is smaller than that in encoding the entire image, and therefore the time required for compression can be shortened.
The occupied amount of the compressed image in the storage device such as AM (Random Access Memory) can be reduced as compared with the case where the entire image is encoded. In addition, since the amount of information in the redundant portion is reduced, the influence on the image quality is small.

An image information amount reducing method for reducing the information amount of a blank portion which is a redundant portion is disclosed in, for example, Japanese Patent Laid-Open No. 2-71.
There is a method disclosed in Japanese Patent No. 670. In this method, pixels with the same layout attribute are connected by a run-length smoothing algorithm, and the pixels subjected to the above processing are subjected to boundary tracking processing to extract a rectangular area, thereby reducing the amount of information in redundant parts. ing.

[0005]

However, the method exemplified above requires the pre-processing by the run-length smoothing algorithm, and the information amount reduction processing can be performed only after the rectangular area is determined in advance. In addition, there is a problem that the processing time is longer than that in the method in which the pre-processing is not performed, and there is a problem that the board occupying area becomes large when the hardware is used. Further, the run-length smoothing algorithm described in the above publication is intended for a black and white binary image and has a problem that it cannot be applied to other gradation images such as a halftone image and a color image.

The present invention has been made in view of the above problems, and the redundant information of an image is reduced by reducing the information amount of a blank portion in the main scanning direction and the sub scanning direction which is a redundant portion in real time. The purpose is to reduce the amount.

[0007]

In order to solve the above-mentioned problems, an image information amount reduction method of the present invention discriminates whether all pixels of each line of an original image are 0 or include at least one non-zero pixel. However, when one or more pixels that are not 0 are included, each line includes an all-zero block in which all the pixels are 0 and one or more pixels that are not 0 and is a natural multiple of a predetermined value in the main scanning direction. It is classified into non-all-zero blocks having a length, the position from the beginning of the line of each non-all-zero block and the length in the main scanning direction are measured, and the above discrimination, classification and measurement are performed for all lines of the original image. Iteratively repeats from the beginning, and based on the results of the above discrimination, classification, and measurement, the original image includes all zero regions in which all the pixels are 0, and one or more pixels that are not 0, and mutually predetermined in the main scanning direction and the sub scanning direction. Divide it into non-all-zero areas that have an interval of more than the number of pixels, and The information of all zero area, and replaces the position of each of the non-all-zero regions of the original image, the size of the main scanning direction and the sub scanning direction.

Further, in the image information amount reducing apparatus of the present invention, an image memory in which an original image whose information amount is to be reduced is stored, and an original image is scanned so that all pixels in the original image are 0. In the original image of each non-all-zero region, the zero-region is divided into a non-all-zero region which includes one or more non-zero pixels and is spaced from each other by a predetermined number of pixels or more in the main scanning direction and the sub-scanning direction. A value extracting unit that extracts the position and the size in the main scanning direction and the sub-scanning direction as each value representing each non-all-zero region,
An image in which image data in each non-all-zero region is extracted from the original image stored in the image memory based on each value extracted by each value extraction unit, and the image data and each value are output. It is equipped with an extraction part, and the information of the blank part is
By replacing each structural block with the position in the original image, the width, and the size in the sub-scanning direction, the redundant information amount of the image is reduced.

[0009]

It is determined whether the pixels of each line of the original image are all 0 or include at least one non-zero pixel, and if each pixel includes at least one non-zero pixel, all the pixels in each line are It is classified into an all-zero block (hereinafter, referred to as a blank block) and a non-all-zero block (hereinafter, referred to as an image block) that includes one or more non-zero pixels and has a length that is an integral multiple of a predetermined value in the main scanning direction. It Next, the number of pixels from the line head of each image block (hereinafter referred to as the head position) and the length in the main scanning direction (hereinafter referred to as the width) are measured. The discrimination, classification, and measurement are sequentially repeated for all lines of the original image from the beginning of the line. From the discrimination, classification, and measurement results, the original image shows all-zero regions (hereinafter referred to as margins) in which all pixels are 0. (Referred to as a portion), which includes one or more non-zero pixels and is divided into non-all-zero regions (hereinafter referred to as structural blocks) at intervals equal to or more than a predetermined number of pixels in the main scanning direction and the sub scanning direction. As a result, the information of the margin portion is replaced with the position (hereinafter simply referred to as position) of each structural block in the original image, the width, and the size in the sub-scanning direction (hereinafter referred to as height), which results in redundant image. The amount of information is reduced.

That is, according to the present invention, the original image includes a blank portion A in which all pixels are 0 and one or more non-zero pixels as shown in FIG. It is divided into a structural block B having an interval equal to or more than a predetermined number of pixels, and the information of the blank area A is replaced with the position, width and height of the structural block B. Therefore, the amount of information in the blank area A can be reduced, and the redundant amount of information in the original image can be compressed.

Since the image data in each structural block B is not coded, if it is coded by a method such as prediction error coding, the compression rate from the original image is the same as when the original image is coded. To improve. When all the pixels of the original image are 0, the code indicating that is the result of the information amount reduction.

Further, the position and width of each structural block B,
Since the height is extracted, it is possible to perform various image processing such as movement, rotation, and deletion on the structural block unit for the image for which the amount of information is reduced according to the present invention.

[0013]

Embodiments of the present invention will be described below with reference to the drawings. In this embodiment, for simplification of description,
Although the original image for which the amount of information is to be reduced is described as a monochrome binary image, it can be applied to a color image.
In the following description, the binary image that is the target of the information amount reduction is composed of white pixels whose pixel values are 0 and black pixels whose pixel values are 1.

FIG. 3 shows an example of a binary image, B
_{1 to} B ₃ are indices of each structural block, 1 _a , 1
_{a + 1} , 1 _b , 1 _{b + 1} , 1 _c , 1 _{c + 1} , 1 _d , and 1 _{d + 1} are the number of lines counted from the leading line (hereinafter simply referred to as “the number of lines”) from p ₁ to p ₃ is the “start position” of each of the structural blocks B _{1 to} B ₃ , and w ₁ to w ₃ are the structural blocks B ₁
The "width" of each ~B _3, h ₁ ~h ₃ is a natural number representing the "height" of each building block B ₁ ~B _3, building blocks B ₁ .about.B ₃ mutually predetermined value It is assumed that they have the above intervals. In this embodiment, the four values indicating the "number of lines", "start position", "width", and "height" are collectively referred to as each value.

FIG. 4 is an enlarged view of the structural block B ₁ of FIG. 3, and FIG. 5 is a diagram showing the 1a _{+ 1} line of FIG.

As can be seen from FIGS. 3, 4 and 5.
The structural block B ₁ (similarly for B ₂ and B ₃ ) is composed of one or more image blocks C, and the image block C is composed of one or more predetermined number of consecutive pixel groups (hereinafter referred to as blocks). There is. In FIGS. 3, 4, and 5, dotted lines represent boundaries between pixels, white rectangular frames represent white pixels, and black painted rectangular frames represent black pixels. Further, in FIG. 3, a thick line represents a boundary between the blank space portion A and the structural block B, and in FIGS. 4 and 5, a thin horizontal line represents a boundary between lines and a thin vertical line represents a boundary between blocks. .. Note that the position of each structural block B may be any of the positions of the pixels at the four corners of each structural block B, but in the present embodiment, the upper left corner of each structural block B is used from the viewpoint of ease of calculation. The pixel positions of are used. In FIG. 3, the structural block B
3, the number C of image blocks is 12 in FIG. 4, and the number of pixels in one block is 4 in FIGS. 4 and 5, but the number is not limited to the above.

FIG. 6 is a flow chart showing the processing procedure of the information amount reduction of this embodiment. In FIG. 6, the value transmission means that the value obtained by subtracting a temporary variable (hereinafter referred to as “auxiliary height”) for calculating the height from the height is the height, and is the “number of lines” of each structural block. The "start position", "width", and "height" (the above four values are collectively referred to as "values") are arranged by the method described below, and each value of each structural block and image data in each structural block are sorted. It means to send out, and the all-zero page code means a code indicating that the values of all pixels in the page are 0.

X is the number of pixels from the beginning of the line, and y
Is the number of lines, w is a temporary variable for calculating the width (hereinafter referred to as “measured width”), and wl is a temporary variable for counting the number of all white lines (hereinafter referred to as “white line number”). ),
wb is a temporary variable (hereinafter referred to as “white block number”) for counting the number of blocks whose pixel values are all white (hereinafter referred to as “white block”), and bp is the structural block group being extracted. Temporary variable that points to the beginning of
, J is a temporary variable indicating the structural block being extracted, and l _j , p _j , w _j , and h _j are the number of lines, the head position, the width, and the height of the j-th structural block, D _j is each value (number of lines, start position, width, height) of the j-th structural block, and H _j is “auxiliary height” of the j-th structural block.
, Flag is a temporary variable (hereinafter referred to as “white block flag”) indicating whether or not all pixels in the block are white, i is a temporary variable indicating the number of repetitions, and LTH is all pixels. BTH is a threshold of the number of consecutive white lines (hereinafter referred to as “line threshold”), and BTH is a threshold of the number of consecutive white blocks (hereinafter referred to as “block threshold”).
BSZ represents a constant representing the number of pixels in one block (hereinafter referred to as “block length”), j = bp substitutes the value of bp in j, and j = j + 1 sets 1 to the value of j. In addition, D _{j + 1} = D _j shifts each value of the j-th structural block to the (j + 1) -th position by max (w,
w _j ) is the larger value of w and w _j , (x, y) is the value of the pixel at the position of the head position x and the number of lines y, and (x, y
y) ≠ 0 represents that the pixel at the position of the head position x and the number of lines y is a black pixel.

The method for reducing the amount of information according to the present invention will be described below with reference to FIGS.
This will be described with reference to FIG.

First, the main routine will be described with reference to FIG.

(I) Initialize the values required to extract the values of each structural block (procedure 1), and perform the processing of (II) and subsequent steps described below.

(II) It is checked in sequence from the first pixel of the first line whether or not it is a black pixel. If the line ends before the black pixel appears, the next line is similarly checked. The above is continued until a black pixel appears (process steps 2 to 6). When a black pixel appears, each value extraction process after (III) described below is performed (processing procedure 10, details are processing procedure 11 and later). In the above processing, 1 is added to the number of pixels x from the beginning of the line each time one pixel is checked (procedure 4), and 1 is added to the number of lines y at the end of each line to set the number of pixels x from the beginning of the line to 1. (Processing procedure 6). If the page is finished before the black pixel appears, the all-zero page code is sent and the process is finished (procedure 7).

Next, each value extraction routine will be described with reference to FIG.

(III) The width w is measured starting from the black pixel appearing in the previous processing (II) (processing procedure 6
0).

As shown in the width measuring routine of FIG. 8, the width w is measured by successively determining the block threshold value BTH with white blocks starting from the black pixel appearing in the previous processing (II).
Check the contents of the block until the number of
This is carried out by adding 1 to the measurement width w each time it is checked (procedures 61 to 72). If the line end occurs while checking the blocks before the number of white blocks successively appear in the block threshold value BTH (procedure 64), 1 is added to the measurement width w at that point to end the width measurement. If the line ends after checking the block (process procedure 73) (process procedure 71), the width measurement is ended without changing the measurement width w. The block threshold BTH is set to determine the interval between the structural blocks in the main scanning direction, and by setting BTH to a certain value,
It is possible to determine how far the pixels in the main scanning direction are included in one structural block.

Next, as shown in processing procedure 11 of FIG. 7, the number of lines and the start position are stored (l _j = y, p _j =
x), the value (width) w _j obtained by multiplying the value obtained by subtracting the block threshold BTH from the measurement width w by the block length BSZ is stored (w _j = (w−BTH) × BSZ), and the number of pixels from the beginning of the line is stored. A value obtained by multiplying x by the measurement width w and the block length BSZ is added (x = x + w × BSZ), and described below (I
V) and subsequent processing is performed.

(IV) If the line has not ended in the previous processing, it is checked in sequence from the pixel at the time of the processing whether the pixel is a black pixel (processing steps 12 to 1).
4) When the black pixel appears, the processing from (III) is performed again. If the line is finished in the previous process, or if the line is finished before black pixels appear (procedure 12,
15), 1 is added to the number of lines y, and the number of pixels x from the beginning of the line is set to 1 (procedure 17), and the structural block extraction process (procedure 20, the details of the procedure 21 and later) is described below (V). ) Is performed, and when the page ends before a black pixel appears (procedure 15), each value is transmitted (procedure 16), and the process ends. In the above process, every time one pixel is checked, 1 is added to the number of pixels x from the beginning of the line (process procedure 13).

Next, the structural block extraction routine will be described with reference to FIG.

(V) Whether or not the pixel is a black pixel is checked in order from the pixel at the time when the previous processing is completed (procedures 21 to 21).
23) When black pixels appear, image block extraction processing (processing procedure 40, details processing procedure 41 and later) after (VI) described below is performed. In the above process, every time one pixel is checked, 1 is added to the number of pixels x from the beginning of the line (process procedure 23). When the line ends before black pixels appear (procedures 22 and 24), 1 is added to the number of lines y, the number of pixels x from the beginning of the line is set to 1, and the number of white lines wl is added to 1 for each structural block. 1 is added to each of the auxiliary heights H _j (processing procedure 25). As a result of the above processing, the number of white lines w
If l is smaller than the line threshold LTH (processing procedure 2
6) Then, the process after (V) is performed again. Line threshold LTH
If it is equal to, each value is transmitted (procedure 27), the number of white lines wl is set to 0, and the auxiliary height H _j of each structural block is set to 0 (procedure 28). Structural block group search processing (processing procedure 50, details of processing procedures 51 to 55) is performed. When the page is finished before the black pixel appears (procedure 24), each value is transmitted (procedure 29), and the process is terminated. The process of (V) is
A routine for extracting structural blocks, where the height of the structural block and the auxiliary height measure the true height of the structural block.
The line threshold LTH is set to determine the interval between the structural blocks in the sub-scanning direction, and LT
By setting H to a certain value, it is possible to determine how far the pixels in the sub-scanning direction are to be included in one structural block.

Next, the image block extraction routine will be described with reference to FIG.

(VI) The width is measured by the same method as in (III), starting with the pixel at the time when the previous processing is completed (procedure 60), and then the positional relationship investigation processing in (VIII) described later ( The processing procedure 80, details of the processing procedure 81 and thereafter, are performed. Further, a value obtained by multiplying the number of pixels x from the head of the line by the measurement width w and the block length BSZ is added (processing procedure 4
1). After the above process, if the line is not finished, it is checked in sequence from the pixel at the time when the above process is completed to see if it is a black pixel (procedures 42 to 44). ) Perform the subsequent processing. In the above process, every time one pixel is checked, 1 is added to the number of pixels x from the beginning of the line (process procedure 43). After the above process,
When the line ends (procedures 42 and 45), 1 is added to the line number y to set the pixel number x from the beginning of the line to 1,
The number of white lines wl is set to 0 (procedure 46), and the process shown in FIG.
Structural block extraction processing after (V) shown in FIG. After the above processing, when the page ends (processing procedure 45),
Each value is transmitted (procedure 47), and the process ends.

Next, the next structural block group search routine will be described with reference to FIG.

(VII) It is checked whether or not the pixel is a black pixel in order from the pixel at the time when the previous process is completed, and if the line ends before the black pixel appears, the next line is similarly checked (process procedure 51). ~ 54). When a black pixel appears, each value extraction process after (III) is performed again (see process procedure 50 in FIG. 9 and process procedure 20 in FIG. 7). In the above processing, 1 is added to the number of pixels x from the beginning of the line each time one pixel is checked (procedure 53), and 1 is added to the number of lines y at the end of each line to set the number of pixels x from the beginning of the line to 1. (Processing procedure 55). If the page ends before the black pixel appears, the process ends without doing anything.

Next, FIG. 1 shows the positional relationship investigation routine.
This will be described with reference to the flowcharts of FIGS. 2 to 16 and the explanatory diagrams of FIGS.

(VIII) From the number of lines y at the time when the image block extraction processing of (VI) is completed, the number of pixels x from the beginning of the line, and the width w measured by the processing procedure 60 of (VI) this time. The positional relationship between the extracted image block and each value that has already been extracted is examined by the following method, and each value is rewritten (process steps 81 to 90).

17A to 24A, dotted lines indicate boundaries between lines, thick lines indicate boundaries between blank areas and image blocks, and shaded areas indicate structural blocks. 17 (b) to 24 (b) show the arrangement state of each value, that is, the number of lines, the head position, the width, and the height, and in FIGS. 17 (b) to 24 (b). , A thick line in the portion representing the arrangement of each value represents the boundary between the values.

(1) As shown in FIG. 17A, the head position p ₂ of the image block is before the head position p ₁ of the structural block whose positional relationship is being checked (p ₂ <p ₁ ), and the image When there is a space equal to or greater than the block threshold BTH between the right end (p ₂ + w ₂ ) of the block and the head position p ₁ of the structural block (p ₁ − (p ₂ + w ₂ )> BTH),
As shown in (b), each value l ₂ , p ₂ , w ₂ , 1 indicating the number of lines, the head position, the width, and the height of this image block.
However, the values are replaced and stored so as to come before the respective values l ₁ , p ₁ , w ₁ , and h ₁ of the structural block (processing procedure 100 in FIG. 12, details are processing procedures 101 and 102 in FIG. 13).

(2) As shown in FIG. 18A, the head position p ₂ of the image block is before the head position p ₁ of the structural block whose positional relationship is being checked (p ₂ <p ₁ ), The right edge (p ₂ + w ₂ ) of the image block is located at a position where the number of pixels from the beginning of the line is smaller than the right edge (p ₁ + w ₁ ) of the structural block (p ₂ + w ₂ <p ₁ + w ₁ ), and the right edge of the image block (p ₂ + w ₂ <p ₁ + w ₁ ). p ₂ + w ₂ ) and the start position p _{1 of the} structural block
If there is a block threshold BTH smaller spacing between the _{(p 1 - (p 2 +} w 2) <BTH) In, and FIG. 18 (b)
As shown in, the difference between the head position p ₂ of the head position p ₁ and the image block of the building blocks, the width w ₁ of the structure block sum (p ₁ -p ₂ + w ₁₎ of the building blocks The width w ₁ ′ is set, and the start position p ₂ of the image block is stored as the start position of the structural block (processing procedure 110 in FIG. 12, details are processing procedures 111, 112, 1 in FIG. 14).
14).

(3) As shown in FIG. 19A, the head position p ₂ of the image block is before the head position p ₁ of the structural block whose positional relationship is being checked (p ₂ <p ₁ ), If the right end (p ₂ + w ₂ ) of the image block is located at a position where the number of pixels from the beginning of the line is larger than the right end (p ₁ + w ₁ ) of the structure block (p ₂ + w ₂ > p ₁ + w ₁ ), then FIG.
As shown in (b), the width w ₂ of the image block is set as the width of the structure block, and the head position p ₂ of the image block is stored as the head position of the structure block (processing procedure 110 in FIG. 12, details are shown in FIG. 14 processing procedures 111, 113-
114).

(4) As shown in FIG. 20A, when the image block has the same value of the head position as the structural block whose positional relationship is being checked (p ₂ = p ₁ ), the width of the image block The larger value (w _{2 in} this case) of w ₂ and the width w ₁ of the structural block is stored as the width of the structural block (processing procedure 120 of FIG. 12, and processing procedure 121 of FIG. 15 for details).

(5) As shown in FIG. 21A, the start position p ₂ of the image block is after the start position p ₁ of the structural block whose positional relationship is being checked (p ₂ > p ₁ ),
When the right end (p ₂ + w ₂ ) of the image block is at a position where the number of pixels from the beginning of the line is smaller than or equal to the right end (p ₁ + w ₁ ) of the structure block (p ₂ + w ₂ ≦ p ₁ +)
As shown in FIG. 21B, the head position and width of the structural block are not rewritten in w ₁ ) (processing procedure 130 of FIG. 12, details of processing procedures 131 and 133 of FIG. 16).

(6) As shown in FIG. 22A, the start position p ₂ of the image block is after the start position p ₁ of the structural block whose positional relationship is being checked (p ₂ > p ₁ ),
There is a space smaller than the block threshold BTH between the head position p ₂ of the image block and the right end (p ₁ + w ₁ ) of the structure block (p ₂ − (p ₁ + w ₁ ) <BTH), and the right end of the image block ( When p ₂ + w ₂ ) is located at a position where the number of pixels from the beginning of the line is larger than the right end (p ₁ + w ₁ ) of the structural block (p ₂ + w ₂ > p ₁ + w ₁ ), FIG.
(B), the head position p ₂ of the image block and the difference between the head position p ₁ of the building block, the sum of the width w ₂ of the image block (p ₂ -p ₁ + w ₂₎ the structure It is stored as the block width w ₁ '(procedure 131 to 13)
3).

(7) As shown in FIG. 23A, the start position p ₂ of the image block is after the start position p ₁ of the structural block whose positional relationship is being checked, and the start position p _{2 of the} image block is And the right edge of the structural block (p ₁ + w ₁ )
When there is an interval of block threshold BTH or more between
_{2 - (p 1 + w 1} ) to ≧ BTH), as shown in FIG. 23 (b), 1 is added to the auxiliary height H _j of the structure block (procedure 88 of FIG. 12), as a result, the auxiliary When the height H _j becomes the same value as the line threshold LTH (procedure 8
9) Then, the values of the structural block are replaced and stored so as to come before the respective values of the structural block having the highest rank (processing procedure 90). The meaning of the rank will be described later.

(8) As shown in FIG. 24A, the head position p ₃ of the image block is located at a position where the number of pixels from the head of the line is larger than the right end (p ₂ + w ₂ ) of the structural block having the lowest rank. In some cases (p ₃ > p ₂ + w ₂ ), FIG.
As shown in FIG. 4 (b), each value of the image block is stored so as to come after each value of the structure block (processing procedure 8).
7).

(9) The above is repeated in order from the structure block having the highest rank until each value of the image block is stored, and the height h _j is the difference (l-1) between the current line number and the line number of the structure block. If it is not larger than _j ), 1 is added to the height h _j of the structural block subjected to the positional relationship investigation (procedure 115 to 116 or procedure 122 to 123 or procedure 134 to 135) and the auxiliary height is set to 0. .. This processing corresponds to the processing of H _j = 0 in the processing procedures 114, 121 and 133.

The above is the information amount reduction method of the present invention.

In the above method, the method of determining the order of the structural blocks is as follows. As shown in FIG. 25, the original image is divided into some structural block groups F with the all-white line E equal to or larger than the line threshold LHT as a delimiter, and the structural blocks are stored in order from the top line. In the structural block group F, the values of the starting position are stored in order from the smallest value, and when the starting positions are the same, they are stored in the order of being closer to the starting line. Note that, in FIG. 25, the thin line indicates a portion having non-zero pixels, the thick line indicates the boundary between the margin portion A and the structural block B, and the dotted line indicates the margin portion A.
And the structural block group F are shown. Note that, here, the portion excluding the all-white line E among the portions obtained by dividing the original image with the all-white line E equal to or larger than the line threshold LHT as a partition is referred to as a structural block group. Therefore, the structural block group includes the left and right margins of the structural block.

FIG. 26 is a diagram in which the structural blocks B in FIG. 25 are given priority in storage according to the above-mentioned order. The thin line in the figure indicates a portion having non-zero pixels, and the thick line indicates a blank portion A. The boundaries with the structural block B are shown respectively. Note that the above-mentioned priority order is not absolute, and may be temporarily changed during the processing depending on the size and arrangement of each structural block, as in the case of (1) or (7) of (VIII) above. However, when sending each value, the above priority order is followed.

Further, the order of the respective values when transmitting the respective values is, for example, as shown in FIG. 27, the position (the number of lines and the head position) of each structural block, the width and the height are predetermined for each structural block. They are arranged in the order of. In FIG. 27, 1 is the number of lines, p is the start position, w is the width,
h represents the height, respectively. The order of the above values may be any order as long as it does not replace the values of different structural blocks.

A method of organizing the values of the structural blocks will be described below with reference to FIGS. 28 to 30. Note that FIG.
In (a) to FIG. 30 (a), a thin line represents a portion having non-zero pixels, and a thick line frame represents a boundary between a blank portion and a structural block. In FIGS. 28 (b) to 30 (b), The thick line in the portion representing the arrangement represents the boundary between each value of each structural block. Also, min (a, b) represents the smaller value of a and b, and max (a, b) represents the larger value of a and b.

[0051] The value obtained by subtracting the auxiliary height H ₁ from the height h ₁ of the building block, and the height h ₁ 'of the building block (see Figure 28).

When the interval S in the main scanning direction between a certain structural block and another structural block is equal to or smaller than the block threshold BTH (see FIGS. 29 and 30), the structural block with the larger value of the head position is detected. The width w ₂ and the width w ₁ of the structural block having the smaller value of the head position and the mutual distance (S = p ₂ −
p ₁ −w ₁ ) and the width of the preceding structural block,
The sum (h ₂ +) of the height h ₂ of the structural block having the larger number of lines (here, the structural block on the right side) of both structural blocks and the difference (l ₂ −l ₁ ) between the numbers of lines of both structural blocks.
l ₂ −l ₁ ) and the height h ₁ of the structural block with the smaller number of lines (here, the structural block on the left side), the larger value is taken as the height h ₁ ′ of the structural block of the previous rank (h
₁ ′ = max (l ₂ −l ₁ + h ₂ , h ₁ ), and the smaller value of the start positions p ₁ and p ₂ of both structure blocks is set as the start position p ₁ ′ of the structure block of the previous rank (p ₁ '= mi
n (p ₁ , p ₂ )), the number of lines l ₁ in both structural blocks,
The smaller value of l ₂ is defined as the number of lines l ₁ ′ of the structural block of the previous rank (l ₁ ′ = min (l ₁ ,
l ₂ )).

The information amount compression method described above is applied to 1 _a = 2, 1 _{a + 1} = 3, 1 _b = 14, 1 _{b + 1} = 1.
5, p ₁ = 2 p ₂ = 21, and when the block length BSZ is 4, the line threshold LTH is 2, and the block threshold BTH is 1 from the first line to the 17th line of the image in FIG. Become.

1. First, when the processing steps 2 to 6 are performed from the head pixel of the head line, a black pixel appears in the third pixel of the third line, and thus the width is measured with this pixel as the head. As a result, the block starting from the 15th pixel is the white block, and the block containing one or more black pixels is three.
Since the measurement width is 4, the width w ₁ of the structural block B ₁ is 12, and this width 12, the number of lines 3 and the head position 3 are stored. At this point, the number of pixels from the beginning of the line is 19
Becomes

2. Next, when the processing steps 12 to 14 are performed, the 25th pixel of the third line is a black pixel, and therefore the width is measured with this pixel as the head. Since the line ends when the measurement of this width ends, the width w ₂ of the structural block B ₂ becomes 8, and this width 8, the number of lines 3, and the head position 25 are stored. Also, since the line is finished after the width is measured, the number of lines is 4 and the number of pixels from the beginning of the line is 1. Here, the next processing is the structural block extraction of the processing procedure 20.

3. Next, when the processing steps 21 to 23 are performed from the head pixel of the fourth line, a black pixel appears in the third pixel of this line, and therefore the width is measured with this pixel as the head. As a result, the block having the seventh pixel as the head is a white block, and the block including one or more black pixels is one, so the measurement width is 2. Based on the value 1 obtained by subtracting the block threshold value from this measurement width by the block length 4 and the number of pixels 3 from the beginning of the line, the existing structural blocks (in this case, structural blocks B ₁ and B ₂ ) are processed by the processing procedure 80 and thereafter. the positional relationship between the rank high structural block is performed in the order (in order from the building block B _1), building blocks B ₁ and (VII
There is a positional relationship in the case of (4) of I) and structural block B ₁
Since the width 8 is larger, only 1 is added to the height h ₁ of the stored structural block B ₁ . At this point, the number of pixels from the beginning of the line is 11.

4. Next, when the processing steps 42 to 44 are performed, a black pixel appears in the twelfth pixel, so the width is measured with this pixel as the head. As a result, the block starting at the 16th pixel is a white block, and the block including one or more black pixels is one, so the measurement width is 2. From the value 1 obtained by subtracting the block threshold value from this measurement width by the block length 4 and the number of pixels 12 from the beginning of the line, the positional relationship with the existing structural blocks is determined in order from the structural block having the higher order by the processing procedure 80 and thereafter. When executed, the structural blocks B ₁ and (VI
Since the positional relationship of the case II) of (6), the width w ₁ of the building block B ₁ represents rewritten to 13. At this point, the number of pixels from the beginning of the line is 20.

5. Next, when the processing steps 42 to 44 are performed, a black pixel appears at the 24th pixel, so that the width is measured with this pixel as the head. As a result, the block starting at the 28th pixel is a white block, and the block including one or more black pixels is one, so the measurement width is 2. From the value 1 obtained by subtracting the block threshold value from this measurement width by the block length 4 and the number of pixels 24 from the beginning of the line, the positional relationship with the existing structural blocks is determined in order from the structural block having the higher order by the processing procedure 80 and thereafter. When executed, the structural blocks B ₂ and (VI
Because of the positional relationship in the case (2) of II), the head position 25 of the structural block B ₂ is rewritten to 24, the width 8 is rewritten to 9, and 1 is added to the height h ₁ . At this point, the number of pixels from the beginning of the line is 32.

6. Next, when the processing steps 42 to 44 are performed,
Since a black pixel appears at the 32nd pixel, the width is measured starting from this pixel. As a result, the line ends after the block having the 32nd pixel as the head, and the number of blocks containing one or more black pixels is one, so the measurement width is 1. From the value 4 obtained by multiplying the measurement width 1 by the block length and the number of pixels 32 from the beginning of the line, if the positional relationship with the existing structural block is performed in order from the structural block having the higher order by the processing procedure 80 and thereafter, the structural block B ₂ And (VIII) in the case of (6), the width 9 of the structural block B ₂ is 12
Can be rewritten as

7. Above 3. ~ 6. The same processing as is performed from the first pixel of the lines from the 5th line to the 14th line.

8. Next, when the processing steps 21 to 23 are performed from the first pixel of the 15th line, all the pixels of the 15th line are white pixels, so 1 is added to the number of white lines, and the auxiliary heights of the structural blocks B ₁ and B ₂ are increased. 1 is added to H ₁ and H ₂ . At this point, the number of white lines is 1, which is smaller than the line threshold, so that the procedure 21 from the first pixel of the 16th line
.. are performed, all the pixels in the 16th line are white pixels, so 1 is added to the number of white lines and 1 is added to the auxiliary heights H ₁ and H ₂ of the structural blocks B ₁ and B ₂ . At this point the number of white lines is 2, which is equal to the line threshold, so
Each value is transmitted, the number of white lines is set to 0, and the auxiliary height of each structural block is set to 0.

Above 1. ~ 8. The state of change of each value in the processing of is as shown in FIGS. 31 and 32. In FIGS. 31 and 32, thick lines indicate structural blocks B ₁ and B.
Represents the boundary between each value of ₂ , each value is the number of lines, the start position,
It is assumed that they are arranged in the order of width and height. In addition, as for B ₁ and (4) in the positional relationship with the existing structural block, the block whose width has been measured is the structural block B ₁ and the above (VI).
It means that there is a positional relationship in the case of (4) of II).

According to the information amount reduction method described above,
Each value of each structural block of the original image is extracted, and the structural block can be extracted from the original image based on each of the above values. In other words, the information amount reduction method replaces the information of the blank portion of the original image with each value of the structural block, and therefore the information amount of the redundant portion can be reduced. In the above description, the block length is a natural number range equal to or less than the number of pixels in one line, and the block threshold value is a range in which a value obtained by multiplying the block threshold value and the number of pixels in the block is a natural number equal to or less than the number of pixels in one line. The line thresholds are arbitrary within a range of natural numbers less than or equal to the number of lines on one page, but extremely large values and extremely small values are not useful from the viewpoint of the gist of the present invention and the processing time. Therefore, for example, CCITT facsimile The optimum value may be determined based on the original image such as the standard image. In particular, the block threshold and line threshold determine the number and size of structural blocks depending on their values, so if you want to reduce the amount of information in the blank space between character parts, for example, set the line threshold more than the number of lines in the blank space between lines. When it is desired to set a small value and extract some lines of the character part as one structural block, an optimum value may be determined, for example, a value larger than the number of lines in the blank space between lines.

An image processing apparatus using the information amount reduction method of the present invention can be configured as shown in FIG. 33, for example.
In FIG. 33, reference numeral 1 is each value extraction unit that extracts each value of each structural block from the original image input by the information amount reduction method of the present invention, and stores the original image in the image memory 2, and 2 is the minimum image 1 page. The image memory 3 having a capacity of 3 minutes extracts image data in each structural block from the image stored in the image memory 2 based on each value of each structural block extracted by each value extracting unit 1, An image extraction unit 4 that outputs data and each value is a control unit that controls the device group.

The original image input to each value extraction unit 1 is stored in the image memory 2 after extracting each value of each structural block. The image extraction unit 3 extracts the image data in each structural block from the original image stored in the image memory 2, and outputs each value and the image data in the order shown in FIG. 34, for example. In FIG. 34, 1, p, w, and h represent the position (the number of lines and the head position), the width, and the height of the first structural block, and the image data 1 represents the image data in the first structural block. The order of the structural blocks follows the above-mentioned order determination method.

In the present embodiment, the original image for which the amount of information is to be reduced is limited to the binary image. However, in the case of a multi-value halftone image, the original image is a blank part in which all pixels are 0 and not 0. A structure block that includes one or more pixels and is spaced from each other by a predetermined number of pixels or more in the main scanning direction and the sub-scanning direction, and the information of the margin is the position and width of each structure block.
The height may be replaced with the image data in each structural block. Further, in the case of a color image, the same processing as that in the case of the multi-value halftone image may be applied to each color of the original image separated into a predetermined color system.

Further, in the present invention, since each value and the image data in each structural block are extracted for each structural block, various image processing is performed for each structural block on the image data after the information amount reduction. It is possible.

[0068]

As described above, according to the present invention,
The original image is divided into a blank part in which all the pixels are 0, and a structural block which includes one or more non-zero pixels and is spaced from each other by a predetermined number of pixels or more in the main scanning direction and the sub scanning direction,
By replacing the information on the margin with the position, width, height, and image data in each structural block, the redundant information amount of the image can be compressed.

[Brief description of drawings]

FIG. 1 is a diagram showing an example of an original image.

FIG. 2 is a diagram showing an example in which an original image is divided into a blank portion and a structural block.

3 is a diagram in which boundaries between pixels, boundaries between blank blocks and structure blocks, indexes of each structure block, and each value of each structure block are added to the original image of FIG.

FIG. 4 is an enlarged view of the structural block B ₂ of FIG.

FIG. 5 is a diagram illustrating the l _{a + 1} line of FIG. 3;

FIG. 6 is a flowchart showing a main routine of processing in this embodiment.

FIG. 7 is a flowchart showing each value extraction routine.

FIG. 8 is a flowchart showing a width measurement routine.

FIG. 9 is a flowchart showing a structural block extraction routine.

FIG. 10 is a flowchart showing an image block extraction routine.

FIG. 11 is a flowchart showing a next structure block group search routine.

FIG. 12 is a flowchart showing a positional relationship investigation routine.

FIG. 13 is a flowchart showing each value change 1 routine.

FIG. 14 is a flowchart showing each value change 2.3 routine.

FIG. 15 is a flowchart showing each value change 4 routine.

FIG. 16 is a flowchart showing each value changing routine 5/6.

FIG. 17 is a diagram showing a positional relationship in the case where the entire image block is located on the left side with respect to the existing structural block and the interval between the blocks is equal to or larger than the block threshold value, and the rewriting of each value in this case.

FIG. 18 is a diagram showing a positional relationship when the entire image block is located on the left side of the existing structural block and the interval between the blocks is smaller than the block threshold value, and the rewriting of each value in this case.

FIG. 19 shows a positional relationship in the case where the left end of the image block is located on the left side of the left end of the existing structural block and the right end of the image block is located on the right side of the right end of the structural block, and of the respective values in this case. It is a figure which shows rewriting.

FIG. 20 is a diagram showing a positional relationship when the left end of an existing structural block matches the left end of an image block and the right end of the image block is located on the right side with respect to the right end of the structural block, and the rewriting of each value in this case. Is.

FIG. 21 shows a positional relationship in the case where the left end of the image block is located on the right side of the left end of the existing structural block and the right end of the image block is located on the left side of the right end of the structural block, and the relationship between each value in this case. It is a figure which shows rewriting.

FIG. 22 is a diagram showing a positional relationship in the case where the entire image block is located on the right side of an existing structure block and an interval between blocks is smaller than a block threshold value, and rewriting of respective values in this case.

FIG. 23 is a diagram showing a positional relationship in the case where the entire image block is located on the right side with respect to the existing structural block and the interval between the blocks is equal to or larger than the block threshold value, and the rewriting of each value in this case.

FIG. 24 is a positional relationship in the case where the entire image block is located on the right side with respect to a plurality of existing structure blocks and the interval between the image block and the adjacent structure block is equal to or larger than the block threshold, and rewriting of each value in this case. FIG.

FIG. 25 is a diagram in which the original image of FIG. 2 is divided into a group of structural blocks and continuous all-white lines whose number is equal to or more than a line threshold.

FIG. 26 is a diagram in which each structural block of the original image of FIG. 2 is ranked.

FIG. 27 is a diagram showing an example of the storage order of each value of each structural block.

[Fig. 28] Fig. 28 is a diagram illustrating a method of organizing each value of a structural block when transmitting each value.

FIG. 29 is a diagram showing a method of organizing each value of each structural block when transmitting each value when the interval between structural blocks is smaller than the block threshold value.

FIG. 30 is a diagram showing a method of organizing each value of each structural block when transmitting each value in the case of another example in which the interval between structural blocks is smaller than the block threshold value.

FIG. 31 is a first part of a diagram showing how each value of a structural block changes when the information amount compression method of the present invention is applied.

FIG. 32 is a second part of the diagram showing how each value of the structure block changes when the information amount compression method of the present invention is applied.

FIG. 33 is a diagram showing a configuration example of the present invention.

FIG. 34 is a diagram showing an example of the order of each value and image data output from the image extraction unit of FIG. 33.

[Explanation of symbols]

1 each value extraction unit, 2 image memory, 3 image extraction unit, 4
Control part, A margin part, B, B _{1 to} B ₃ structure block, C image block, E all white line, F structure block group, start position of each p _{1 to} p ₃ structure block, w ₁ to
w ₃ width of each structure block, h _{1 to} h ₃ width of each structure block

Claims (2)

[Claims] 1. The pixels of each line of the original image are all 0 or 0.
If one or more non-zero pixels are included, and if one or more non-zero pixels are included, each line includes all zero blocks in which all pixels are 0 and non-zero pixels in 1
It is classified as a non-totally zero block having a length that is a natural number multiple of a predetermined value in the main scanning direction including one or more, and the position from the line head of each non-totally zero block and the length in the main scanning direction are measured, The discrimination, classification, and measurement are sequentially repeated for all lines of the original image from the beginning of the line. From the discrimination, classification, and measurement results, the original image has all-zero regions in which all pixels are 0 and non-zero pixels are 1 If there are two or more of them, they are divided into a non-total zero region having a predetermined number of pixels or more in the main scanning direction and the sub-scanning direction, and the information of the above-mentioned all-zero region is the position in the original image of each non-total-zero region. And the size in the main scanning direction and the size in the sub scanning direction. 2. An image memory in which an original image whose information amount is to be reduced is stored, an all-zero region in which the original image is scanned and all the pixels are 0, and at least one non-zero pixel is included. The main scanning direction and the sub-scanning direction are divided into non-total zero areas having a predetermined number of pixels or more, and the positions of the non-total zero areas in the original image and the sizes in the main scanning direction and the sub-scanning direction. Each value extraction unit for extracting as a value representing each non-total zero region, and each non-total zero from the original image stored in the image memory based on each value extracted by each value extraction unit. An information amount reducing apparatus comprising: an image extracting unit that extracts image data in a region and outputs the image data and the above-mentioned values.