JPS6365189B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image processing method for determining a halftone image area in an image containing a mixture of two-tone images such as characters and graphics and halftone images such as photographs.

Visually good image quality can be expected when characters and figures are expressed in two gradations of black and white, but the information contained in images is lost unless they are expressed in intermediate tones, such as photographs. When a two-tone image area (hereinafter referred to as a character area) and a halftone image area (hereinafter referred to as a photographic area) coexist, generally the entire screen is expressed in halftones, or a ni, 2
Halftones are expressed in a pseudo manner using gradation, but in the former case, the amount of data required for image storage and transmission increases, and in the latter case, the high quality of high spatial frequency parts in the character area is impaired.

Therefore, it is desirable to express the text area in two gradations, and to express the photo area in halftones such as dithering.

The present invention determines a photographic area based on a dithered image, and its purpose is to allow character areas to be expressed in two gradations and photographic areas to be expressed in halftones using the dither method or other methods.

Hereinafter, the first and second image processing methods according to the present invention will be explained in detail.

In the first image processing method of the present invention, the first
As a step, the image is converted into a dithered image by dithering.

As a second step, if the dot pattern of the pixel block corresponding to the dither matrix has low similarity or high similarity between adjacent pixel blocks in the character area, the dot pattern is similar to the background color of the image. In addition, in the photographic area, the photographicity of each part of the dithered image is measured by utilizing the high similarity between adjacent pixel blocks. To measure the photo-likeness, a group of pixels having a size that is an integral multiple of the pixel block of the dither matrix is used as a cell, and the similarity of dot patterns between pixel blocks is measured for each cell. Determine as.

In the next step, that is, the third step, multiple formats of adjacent cell sets are sequentially set and the cell mode image is raster-scanned in a format corresponding to each format of adjacent cell sets, and the adjacent cell set in each focused cell is set in H mode. By converting each cell of interest in non-H mode to H mode under certain conditions, H
The part where mode cells are gathered is converted into a substantially rectangular shape.

As the final step, that is, the fourth step, the size of the approximately rectangular area where the H-mode cells are gathered is measured,
By converting H mode cells in an area smaller than a set size to non-H mode,
Convert H mode cells in the character area to non-H mode.

Furthermore, in the second image processing method of the present invention,
Immediately after the second step described above, a cell block is obtained by integrating a plurality of cells two-dimensionally, and the cell mode image is divided using this cell block as a unit. The processing speed can also be increased by determining the H mode according to the amount of H mode, that is, by executing the processing from the third stage described above on the cell block mode image.

FIG. 1 is an explanatory diagram of a dithered image, and FIG. 2 is a functional block diagram showing an example of an image processing apparatus according to the present invention. Hereinafter, an embodiment of the present invention relating to the above-mentioned second image processing method will be described in detail with reference to FIGS. 1 and 2.

The dither matrix shown in Fig. 1 is one type of 17-gradation display, and the gradation of the original image A is judged binary using the corresponding threshold value of such a dither matrix (DM). This results in a dithered image B.

In FIG. 1, the numbers attached to the pixels of the original image A indicate the gradation order of each pixel in the original, with 0 being the lowest density and 15 being the highest density. The numbers 0 to 15 in the dither matrix (DM) indicate an example of the order of threshold values, and the binary values 0 and 1 in the dither image B indicate the bit pattern of the dither image.

In FIG. 2, 1 is an input terminal for dithered images;
2 is a dithered image memory that stores one screen of dithered images; 3 is an input terminal for a two-gradation image obtained by converting the same original image into two gradations using a uniformly set threshold value; and 4 is an input terminal for a two-gradation image. This is a two-tone image memory that stores . Both images are 8 pixels/mm in both main and sub-scanning directions.
It has a pixel density of approximately

The cell determination unit 5 selects a cell (2
The dithered image plane is divided into sections (equivalent to 2mm x 2mm) and the photo-likeness of each cell is determined.First, the topmost 1 pixel row of each cell, that is, a 16-pixel dot pattern, is stored in the dithered image memory. If the dot patterns corresponding to the pixel blocks satisfy the following conditions a and b, then the cell has a photographic quality. As,H
Decide on mode.

Condition a: The 4-bit dot pattern corresponding to each pixel block is the same for all 4 pixel blocks.

Condition b: The 4-bit dot pattern is (0,
0, 0, 0).

FIG. 3 is a diagram shown to explain the relationship between a dither image and a cell. In cell A, all 4-bit dot patterns corresponding to pixel blocks are (1,
0, 1, 0) and this is (0, 0, 0,
0), the above condition is satisfied and the H mode is determined to be "1", and since the dot pattern in cell B is different, it is determined to be the non-H mode "0". In this manner, a determination is made for each cell, and the results are stored in the cell mode memory 6 in FIG.

FIG. 4 shows an example of a cell mode image in the cell mode memory 6, where the area indicated by two short line segments corresponds to one H mode cell, and the H mode cells are concentrated. Two areas C ₁ and C ₂
corresponds to a photo area, and the other parts are character areas, but it shows that cells determined to be in H mode remain. If the cell size setting is increased, non-H mode cells will increase in the photo area, so a relatively small value as in this example is desirable.

Note that when making a determination using a small number of representative pixels in each cell as in this cell determination section 5, if pixel rows or pixel columns whose threshold values in the dither matrix are significantly different are selected, The mode decision may be made with little regard to the position within the cell, such as top, middle, bottom, etc. within the cell.

In addition, this cell determination unit 5 determines the similarity of dot patterns between pixel blocks based on coincidence, but the altitude is, for example, 10
It is sufficiently practical for several gradations.

In addition, this cell determination unit 5 uses a dot pattern (0, 0,
0, 0), but if it is expected that there will be a large variation in the background color, it is better to include a dot pattern close to the background color to determine the non-H mode. is desirable.

The cell block determination unit 7 in FIG. 2 receives the cell image in the cell mode memory 6 and determines whether each cell block is in the H mode or not, and divides the cells into two cells in both the main and sub-scanning directions. Dimensionally integrated cell block (4mm x
4 mm equivalent), divide the cell mode image plane into cell blocks, take out the number of H mode cells in each cell block, that is, the H mode amount, and set this H mode amount to a preset value ( In this example, this setting value is “2”
shall be. ), that is, the cell block is determined to be in H mode on the condition that the number of H mode cells in the cell block is two or more,
Stored in cell block mode memory 8. The size of the cell block is based on the assumption that there is a distance of about 4 mm or more between the text and the photo, and if the distance is large, a larger size can be set. In this way, by integrating 2×2 cells into one cell, the amount of information to be processed in subsequent stages is reduced to 1/4, reducing processing time, and
The number of H mode cell blocks in the character area is reduced.

FIG. 5 shows a cell block image in the cell block mode memory 8, with four short line segments indicating one cell block in the H mode. As is clear from Figure 5, the photographic area
Not all cell blocks of C ₃ and C ₄ are determined to be in H mode, and some H mode cell blocks remain even in the character area.

The area determination unit 9 in FIG. 2 expands the cell block mode image in the cell block mode memory 8 so that the area where the H mode cell blocks are assembled approaches a rectangle, thereby converting almost all the cell blocks in the photographic area into an H image. mode, and if the enlarged area is smaller than the specified size, the H mode cell block remains in the character area by converting the H mode cell block to the non-H mode. Try not to.

FIG. 6 is an explanatory diagram of rectangular enlargement processing in the area determination section 9.

First, as shown in A-a in FIG.
A set of cell blocks CB1 to CB3 adjacent to the left side is set, and those cell blocks CB1 to CB3 are set to H.
E-b in Figure 6 on the condition that the mode is
As shown in the figure, the cell block of interest in non-H mode is
Convert CB0 to H mode and write to cell block mode memory 8. The area determination section 9 performs this process by shifting the screen of the cell block memory 8 one cell block at a time from left to right, and then by raster scanning by shifting the screen one cell block at a time from top to bottom. When one screen has finished processing,
As shown in row a of FIG. 6, the area determination unit 9
Set adjacent cell block sets CB1 to CB3 adjacent to the left side, lower left side, and lower side, and convert the target cell block CB0 in non-H mode to H mode on the condition that these cell block sets CB1 to CB3 are in H mode. , the process of writing to the cell block memory 8 is performed by raster scanning from the bottom to the top of the screen and from left to right.

Next, the area determining unit 9 determines the lower side, lower right side, and
Adjacent cell block groups CB1 to CB adjacent to the right side
3, and as shown in Knee-a in FIG.
Similar processing is performed by raster scanning corresponding to each type of adjacent cell block sets CB1 to CB3.

In this way, the area determining section 9 sets four types of adjacent cell block sets and repeats twice the process of converting a non-H mode cell block to an H mode. Normally, if adjacent cell block groups are set five times, such as E-a, R-a, H-a, Knee-a, E-a (see Figure 6), the rectangular area will hardly expand; If the settings are made 8 times as shown above, almost all cell blocks in the photo area will be in the H mode, regardless of the quality of the photo.

Figure 7 shows an example of a cell block mode image obtained by repeatedly setting four types of adjacent cell block sets twice.
It shows that C ₅ and C ₆ are expressed by rectangular H-mode cell block regions. Furthermore, photography area
The area indicated by vertical lines in C ₅ and C ₆ is the cell block before processing, and the black area is the enlarged cell block area. As a final step, the area determination unit 9
For this, a 5 x 5 cell block area (20 mm x 20 mm) is set as the minimum size of the photo area, and among the connected cell block areas of H mode obtained in the previous processing, the cell block area is smaller than this size. It is regarded as a misjudgment and converted to non-H mode. By this process, in the cell block mode memory 8, the H mode cell blocks scattered in the character area shown in FIG. 7 are converted to the non-H mode.

The image synthesis section 10 in FIG. 2 takes out the mode information of each cell block stored in the cell block mode memory 8, and if the cell block is in H mode, takes out the dithered image at the corresponding position from the dithered image memory 2. If the cell block is in the non-H mode, it is output from the two-tone image memory 4 at the corresponding position and output to the output terminal 11.

In this embodiment, when determining the H mode for each cell, AND processing is performed for each 16 pixel in the main scanning direction, and majority logic is used for each cell block, so that the result obtained by the above processing is There is a possibility that the set of H-mode cells that are displayed are determined to be narrower on the left and right sides than the original intermediate adjustment area, and in the same way in the sub-scanning direction, the topmost row of 16 pixel rows in the cell is used as the determination row. Furthermore, since majority logic is used in the cell block, there is a possibility that the upper and lower regions are each narrowly determined, so it is desirable to further add the following processing. That is, after converting the H mode cell block area smaller than the specified size to the non-H mode, the cell block mode image in the cell block mode memory 8 is raster scanned to convert the non-H mode cell block directly adjacent to the H mode cell block. It is recommended to convert the image to H mode and enlarge each photographic area by one cell block in the top, bottom, left and right directions.

As explained above, in this embodiment, it is first determined for each cell whether or not it is a halftone region based on the distribution state of white and black pixels of the dithered image, and then based on this, the halftone region is further determined for each cell block. Since the determination is made as to whether it is in the key area, the amount of data to be stored during the determination is small, which simplifies the circuit. Also, from small area (cell) to large area (cell block)
This has the advantage of reducing false judgments compared to configuring cells in large areas.

The embodiment described above is a case where the above-described second image processing method of the present invention is used.

Next, an example of the case where the above-described first image processing method of the present invention is used is as follows.

In other words, if the properties of the original image are roughly determined and it is possible to predict the dot patterns of each pixel block within the cell when determining the H mode of the cell, the range of dot pattern similarity measurement can be set relatively. By setting it wide, the cell size can be increased, and in such a case, the integration into the cell block can be omitted and the process of rectangular expansion can be performed immediately.

As is clear from the above description, the present invention can extract halftone areas and display and transmit images according to the content of the image, so it can be applied to facsimile. Furthermore, when editing an image, conventionally a light pen was used to indicate the position on the CRT display tube, but by applying the present invention, area division is automatically performed, improving operability.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram of a dither image, Fig. 2 is a functional block diagram showing an example of an image processing device according to the present invention, Fig. 3 is an explanatory diagram of mode determination, and Fig. 4 is an example of a cell mode image. 5 is a diagram showing an example of a cell block mode image, FIG. 6 is an explanatory diagram of an example of rectangular enlargement, and FIG. 7 is a diagram showing an example of a cell block mode image obtained by rectangular enlargement. . 1... Dither image input terminal, 2... Dither image memory, 3... Input terminal, 4... 2-gradation image memory, 5... Cell determination section, 6... Cell mode memory, 7... Cell block determination. section, 8... cell block mode memory, 9... area determination section, 10...
Image composition section, 11...output terminal.

Claims (1)

[Claims] 1. A first step of converting an image into a dither image, and converting the dither image surface in cell units using pixel groups having a size that is an integral multiple of the pixel block of the dither matrix in both the main and sub-scanning directions. A dot pattern of pixel blocks in a cell is divided into at least one row of pixel groups continuous in the main scanning direction within the cell and at least one column of pixel groups continuous in the sub-scanning direction within the cell. a second step of determining the cell as H mode on the condition that the dot patterns in the cell are similar to each other and that the dot pattern is not similar to the dot pattern of the ground color; A cell mode image is raster-scanned by sequentially setting adjacent cell sets in multiple formats and in a format corresponding to each adjacent cell set, and the cell of interest in non-H mode is scanned in H mode on the condition that the adjacent cell set is in H mode. mode, and a fourth step of converting each cell in the set to a non-H mode, provided that the set of H-mode cells is less than a set size. Featured image processing method. 2 The first stage of converting an image into a dithered image is to divide the dithered image plane in units of cells using pixel groups whose size is an integral multiple of the pixel block of the dither matrix in both the main and sub-scanning directions, and to divide the dithered image plane into cells. The mutual similarity of dot patterns of pixel blocks in a cell is calculated for at least one of a group of pixels in at least one row continuous in the main scanning direction and a group of pixels in at least one column continuous in the sub-scanning direction within the cell. a second step of determining the cell as H mode on the condition that the dot patterns within the cell are similar to each other and that the dot pattern is not similar to the dot pattern of the ground color; A two-dimensional combination is used as a cell block, and the cell mode image plane is divided into cell blocks, and the amount of H mode expressed by the number of H mode cells in each cell block is measured for each cell block. The third step is to determine each cell block as H mode on the condition that it is the mode amount, and to sequentially set a plurality of adjacent cell block sets and raster scan the cell block mode image in a format corresponding to each adjacent cell block set. , a fourth step of converting the cell block of interest in non-H mode to H mode on the condition that the set of adjacent cell blocks is in H mode, and a fourth step on the condition that the set of H mode cell blocks does not reach a set size. and a fifth step of converting each cell block of the set into a non-H mode.