JP3157861B2 - Area division method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for dividing an intermediate area of a photograph, a graphic image, or the like, and particularly to an apparatus for reading a set original as an image, such as a copying machine, a facsimile, and an OCR. The present invention relates to an area dividing method for dividing an area such as a halftone area and a character area when the halftone area exists and performing separate processing.

[0002]

2. Description of the Related Art Conventionally, as a method of dividing a halftone region, there is a method of binarizing an image and processing a portion having a high luminance value as a halftone region and a portion having a low luminance value as a character region. However, in this method, a process similar to that of a character area is performed on a blackish photograph or the like having a relatively low luminance value, and a flat process without a three-dimensional effect may be performed.

[0003]

SUMMARY OF THE INVENTION The present invention has been made in order to solve such a conventional problem, and is intended to reliably extract a region such as a photograph or a figure having a halftone luminance value. It is intended to provide a division method.

[0004]

An area dividing method according to the present invention comprises the steps of: calculating a difference between luminance values within a predetermined convolution;
The halftone pixel is extracted based on the lower limit value. According to the area dividing method according to the present invention, the halftone area can be reliably extracted.

[0005]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment of an area dividing method according to the present invention will be described with reference to the drawings. In the flowchart of FIG. 1, first, a document is read from an input means such as a scanner (step 1-1), and the document is reduced to an appropriate size (1).
-2). With this size reduction, the dither image is shaded, and the dithered photograph is also processed as a halftone area. Further, the memory capacity required for the subsequent processing can be reduced, and the effect of noise removal by smoothing can be obtained.

Here, the reduction rate is 400 dots / i.
For images input with nch, the size ratio is 1/8 (50 dots)
s / inch) or less, the dither image of about 65 lines can be reliably shaded, and practically 1/16
(25 dots / inch) is preferable. At the time of reduction, a region of a fixed size of the original image is generally represented by one pixel, but the average density of the region is given to the representative pixel in the sense of maximally storing the density information of the entire image. .

The reduced image is binarized (step 1-
3) The region around the halftone region is clearly separated by expansion and contraction processing (step 1-4). There is a possibility that characters such as commentary attached to the vicinity of the photograph may be determined to belong to the region of the photograph. In this case, halftone processing is performed on the character, and the character is blurred. Therefore, such a character area is erased by performing expansion giving priority to high luminance, and further contracted to return to the original area size. This clearly separates the low-luminance photographic areas. At this time, the effect of group reduction and noise removal by the graphic fusion can be obtained.

Next, it is determined whether or not the center pixel is a halftone pixel for each 3 × 3 convolution of the reduced image obtained in step 1-2 (steps 1-5 to 1-8). Since it is necessary to consider not only the density of the pixel but also the density distribution as to whether or not the pixel is a halftone pixel, a convolution process is necessary. In step 1-5, it is determined whether or not the difference between the maximum luminance and the minimum luminance of the pixels in the convolution, that is, the luminance difference is equal to or less than a predetermined value. As shown in FIGS. 5 and 6, a halftone area such as a photograph has a gentle change in luminance, and the luminance difference is a powerful index.

In steps 1-6 and 1-7, the upper and lower limits of the luminance distribution are evaluated, respectively. That is, the center pixel of the convolution whose upper limit is equal to or more than the predetermined value is excluded from the halftone pixels, and the white area of the background is removed. Also, the central pixel of the convolution whose lower limit is equal to or less than a predetermined value is excluded from the halftone area, and a large black character area and the like are excluded. Only pixels that meet the conditions of steps 1-5 to 1-7 are extracted as halftone pixels (step 1-8), and then, for each region of the binary image obtained in step 1-4, halftone pixels Count. A unit area in which the ratio of the count value to the total number of pixels in this area is equal to or greater than a predetermined value is determined as a halftone candidate area (step 1-9).

Next, for a region (FIGS. 2A, 2B, and 2C) determined as a halftone region such as a photograph or a figure, a minimum rectangle circumscribing the region is generated (step 1-1).
0) (FIGS. 3A, 3B, and 3C). In this way, by handling the area in units of rectangles, high-speed memory control is facilitated, and the overall processing can be speeded up. Next, the sizes of two adjacent sides of the rectangle generated in step 1-10 are obtained, and it is determined whether each of them is equal to or smaller than a predetermined value (step 1-11). Here, if the size of one side is equal to or less than a predetermined value, it is determined that the size of a normal photograph, figure, or the like is too small, and is excluded.
Return to It should be noted that a process of excluding a region from the halftone region when the vertical and horizontal proportions of the region are abnormal values may be appropriately adopted.

In step 1-12, when there are a plurality of rectangles, it is determined whether the rectangles partially overlap each other. As shown in FIG. 3, rectangles (A) and (B) partially overlap. In this case, in step 1-13, both rectangles are integrated to form a rectangle (D). This is because it is usually better to judge halftone areas that are adjacent to each other so that rectangles overlap each other as the same photo, figure, etc.
This is because the region can be caught without omission, and the region can be simply divided, thereby speeding up the processing and obtaining a beautiful processing result. Steps 1-12 until a plurality of rectangles no longer overlap
Is repeated, and if duplication is eliminated, step 1
-14.

In step 1-14, the characteristic pixel areas of photographs, figures, etc. are obtained for all the divided rectangles, and it is determined whether or not the area value is equal to or smaller than a predetermined value. Here, if the area value is equal to or larger than the predetermined value, the process ends. If it is found that the area of the feature pixel such as a photograph or a figure in the divided rectangle is smaller than a predetermined value, there is a possibility that the photograph, the figure or the like is arranged as shown in FIG. In this case, for example, it is preferable to alert the user to that effect and eliminate the division process (step 1-15). As a result, it is possible to reliably divide a halftone area such as a photograph or a graphic image from an image area composed of characters and a background.

[0013]

As described above, according to the present invention, when a halftone area such as a photograph or a graphic image exists in a copying machine or the like that reads a set original as an image, the area is composed of characters and a background. This has the effect that the image can be reliably divided from the image area.

[Brief description of the drawings]

FIG. 1 is a flowchart of steps 1-1 to 1-8 showing one embodiment of the present invention.

FIG. 2 is a conceptual diagram showing an area defined as a photograph portion.

FIG. 3 is a diagram for explaining a rectangle.

FIG. 4 is a diagram illustrating the integration of overlapping rectangles.

FIG. 5 is a conceptual diagram illustrating a luminance distribution of an image including a character and a background.

FIG. 6 is a conceptual diagram illustrating a luminance distribution of an image including a photograph, a figure, and the like.

FIG. 7 is a conceptual diagram illustrating an upper limit and a lower limit.

FIG. 8 is a conceptual diagram showing a rectangle when a photograph, a figure, and the like are specially arranged.

FIG. 9 is a flowchart of steps 1-9 to 1-15.

Continuation of front page (72) Inventor Shiro Inoda 22-22, Nagaike-cho, Abeno-ku, Osaka-shi Inside Sharpe Co., Ltd. (56) References JP-A-2-84879 (JP, A) JP-A-3-16380 (JP, A) JP-A-3-141774 (JP, A) JP-A-4-356869 (JP, A) JP-A-64-61170 (JP, A) JP-A-56-116182 (JP, A) JP-A-1 −168167 (JP, A)

Claims (7)

(57) [Claims] An image is subjected to a binarization process to obtain a background image.
Defining each area that is not an image area;
Within a given convolution of a given number of pixels vertically and horizontally
The difference between the brightness values is equal to or less than a predetermined value and the convolution
The maximum luminance value in the application is less than the specified value, and the minimum luminance value is specified.
Value is greater than or equal to
Making the pixel a halftone pixel;
For each area, the number of the halftone pixels included in the area is
Counting, and the above halftone with respect to the number of constituent pixels in the area.
A region where the ratio of the number of pixels is equal to or greater than a predetermined value is a halftone region.
A region dividing method characterized by including the following steps:
Law. Wherein a reduced image, and performs the extraction of the halftone pixel on the reduced image, the area dividing method according to claim 1. 3. The method according to claim 2, wherein after the binarization processing, the image is expanded and
Perform contraction processing to determine each area that is not a background image area
By doing so, the character area around each area is
2. The method according to claim 1, wherein the area is excluded . 4. The area dividing method according to claim 1, wherein a minimum rectangle circumscribing a figure obtained by integrating adjacent halftone areas is regarded as one halftone area. When 5. A plurality of rectangular partially overlapping, the minimum length that circumscribes these rectangles was integrated graphic
5. The area dividing method according to claim 4, wherein the rectangle is regarded as one halftone area. 6. The method according to claim 4, wherein when at least one side of the rectangle is equal to or smaller than a predetermined value, the rectangle is regarded as not being a halftone region. 7. The area dividing method according to claim 4, wherein when the vertical and horizontal proportions of the rectangle deviate from predetermined values, the rectangle is regarded as not a halftone area.