JP4437621B2 - Region detection apparatus and method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a digital copying machine, a facsimile, and other image quality improvement processing of an output image in digital image processing, and more particularly to an area detection technique such as a character / line drawing of an image.
[0002]
[Prior art]
As a conventional technique related to image quality improvement processing, for example, a method described in Japanese Patent Application Laid-Open No. 63-240175 is known. This is because black character areas are detected from the process of detecting character edges using the continuity of pixels and the process of determining black parts, and the detected areas are processed to improve image quality. Is to do. However, such a method for detecting a character edge is particularly strong in halftone dot printing, but there are many areas in the pattern that are erroneously detected as character areas, and this erroneous detection also causes image quality degradation. Yes.
[0003]
In order to reduce this inconvenience, for example, the paper “Image Area Separation Method for Mixed Image of Character / Picture (Half Dot, Photo)” (Electronic Information and Communication Society Transactions Vol. J75-DII No. 1 pp. 39-47 1992 1 (Moon) proposes detection of a halftone dot region in addition to detection of a character edge region, and includes a mechanism that does not determine a halftone dot pattern as a character edge. Japanese Patent Laid-Open Nos. Hei 4-14378 and Hei 7-203198 aim to reduce false detection areas in a pattern by limiting characters to be detected only to thin characters. In the method of detecting only such thin characters, the false separation of the pattern is certainly drastically reduced. On the other hand, from the feature of the algorithm, to detect fairly thick characters, a corresponding large pattern is prepared. There must be. To prepare a large pattern means that a very large number of line buffers are required in the sub-scanning direction when performing detection.
[0004]
In recent years, for example, multi-level control corresponding to the line width as seen in Japanese Patent Laid-Open No. 7-203198 has been performed. However, as the thick character side increases from thin characters to thick characters, the higher the accuracy, the higher the accuracy. It can be said that the more the multi-stage control is performed, the more patterns for calculating the character width have to be increased.
[0005]
[Problems to be solved by the invention]
Thus, in the prior art, if it is possible to detect even a relatively large character, it is necessary to increase a large number of line buffers and patterns for character width calculation. For example, to detect double line width, the line buffer is simply doubled. FIG. 6 shows this state. Regardless of the detection method, it can be seen that B needs to be approximately twice as large as A in order to detect the line width. Further, in order to output a signal for multi-stage control corresponding to the character width, a pattern for stages is simply prepared. FIG. 7 shows this state. This example is an example of generating a four-step control signal, and it can be seen that four patterns are required regardless of the detection method.
[0006]
An object of the present invention is to make it possible to realize character area detection or the like based on line width information without a large increase in hardware even if the character width to be detected increases.
It is another object of the present invention to prevent detection of dense thin line character portions and to easily generate a signal for multistage control.
[0009]
The present invention inputs the image data, in the area detecting device for detecting a character-line region to the input image data, first means for detecting the outline of a character-line drawing of the input image data, the input image A second means for detecting a region corresponding to a line width of two or more different characters / line images, for data and at least one low-resolution image data obtained by converting the resolution of the input image data; When the detection result of the first means indicates a contour, a value determined according to a predetermined table based on the detection result of the second means, and when the detection result of the first means indicates a non-contour And a third means for outputting a predetermined value as a multi-stage determination signal .
[0010]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be specifically described below with reference to the drawings.
[Example 1]
FIG. 1 is a block diagram of Embodiment 1 of the present invention. In FIG. 1, 1 is an image input device such as a scanner, 2 is a resolution conversion circuit (1) that converts an input image signal to a predetermined resolution (to the lower resolution side), and 3 is an edge detection that detects a character / line image edge region. Circuit 4 is a line width detection circuit that determines whether or not the target portion is below a certain line width, and 5 is a resolution conversion circuit that converts the determination result of the line width detection circuit 4 to a predetermined resolution (to the high resolution side). (2) and 6 are AND circuits in which the active portions of both the resolution conversion circuit 5 and the edge detection circuit 3 are the true active portions.
[0011]
The image input device 1 is a scanner using, for example, a CCD element, and after reading an original image and performing A / D conversion processing, outputs an image signal (image data) of approximately 600 dpi 6 bits and a linear density for each pixel. . Here, for the purpose of explanation, it is assumed that “0” represents the white of the original and “63” represents the black of the original as the meaning of the signal. If the input image signal is a color image signal, for example, an RGB signal, the G signal may be used as a representative. Of course, the input image signal may be obtained through a network.
[0012]
The resolution conversion circuit (1) 2 converts the input image signal into an appropriate low resolution signal according to the size of the character / line image (hereinafter referred to as character) to be detected, that is, the line width and the prepared line width pattern. Convert. As a conversion means, a method using convolution has good performance, but thinning out can be easily performed. Further, since the amount of the line buffer is related to the sub-scanning direction, the main scanning may be performed only in the sub-scanning without performing resolution conversion. Here, the image data is converted to 300 dpi by simple thinning in both the main and sub directions.
[0013]
The edge detection circuit 3 is a circuit that detects a region corresponding to a character edge from an input image signal. Conventionally, many methods and apparatuses have been proposed for this type of edge detection, and since edge detection itself is not the gist of the present invention, a detailed description thereof will be omitted. The output of the edge detection circuit 3 is 1 pixel per 1 bit, and the detected edge area pixel is made active (logic 1).
[0014]
The line width detection circuit 4 is a circuit that determines whether or not the target or target block is within a predetermined area having a certain width. This is a circuit for detecting whether or not. For example, it is possible to apply a white background detection circuit described in Japanese Patent Application Laid-Open No. 4-248766 and determine whether or not a predetermined character width is based on the presence of a white background in a predetermined block (such as 5 × 5). The output of the line width detection circuit 4 is also 1 pixel per 1 bit, and a pixel within a predetermined (character) width is activated (logic 1). Here, as long as the resolution of the original image is the same as the data resolution before detection of the white background is low, even large-size characters can be detected.
[0015]
The resolution conversion circuit (2) 5 is a circuit that reverses the resolution conversion circuit (1) 2 and returns the result from the line width detection circuit 4 to the original resolution. Here, a simple double writing process is performed in both the main and sub directions to convert the image data to 600 dpi.
[0016]
In the AND circuit 6, a pixel in which an output from the resolution conversion circuit (2) 5 is active and an output from the edge detection circuit 3 is also active is set to active (logic 1) as a true character area.
[0017]
[Example 2]
In the first embodiment, there is a case where the character cannot be detected as a character although it is desired to detect the character by the conversion rate executed by the resolution conversion circuit (1) 2 and a pattern in which lines are densely arranged at a certain interval. For example, when the lines arranged at intervals of one pixel as shown in FIG. 4 are thinned out for each pixel, there is a possibility that the lines become solid black. The second embodiment has a mechanism that avoids such problems.
[0018]
FIG. 2 is a block diagram of the second embodiment of the present invention, where 1 is an image input device such as a scanner (of course, an image may be captured from a network), and 2 is an input signal having a predetermined resolution (to the lower resolution side). Resolution conversion circuit (1) for conversion, 3 is an edge detection circuit for detecting a character edge region, 4-1 and 4-2 are line width detection circuits for determining whether or not the target portion is below a certain line width, 5 Is a resolution conversion circuit (2) for converting the determination result of the line width detection circuit 4-1 into a predetermined resolution (to the high resolution side), and 6 is a true active part of both the OR circuit 7 and the edge detection circuit 3 And an AND circuit. Here, the OR circuit 7 is an OR circuit that takes a logical sum of the results of the line width detection circuit 4-2 and the resolution conversion circuit (2) 5.
[0019]
The differences from the first embodiment will be described below. In the second embodiment, line width detection circuits 4-1 and 4-2 perform line width detection on the original image data and on the image data after resolution conversion (50%), respectively. The result is added by the OR circuit 7. Thereby, different line widths can be detected by the line width detection circuits 4-1 and 4-2 having the same configuration.
[0020]
Of course, the parameter corresponding to the line width of the line width detection circuit may be different depending on the purpose (for example, the upper limit of the line width to be taken or the dense pattern previously). In addition, two types of resolution data (same size and 50%) are performed. By increasing this type, more (larger size) characters can be used as characters, and more dense lines can be obtained. Many types of patterns can be detected as characters.
[0021]
[Example 3]
For example, in image processing such as spatial filter processing and UCR processing, there are cases in which multistage control is desired in addition to 0/1 control. In other words, if you want to perform strong edge enhancement processing as the character width becomes thinner, or if the black character width becomes thinner, the black ratio (ratio that replaces the minimum density of yellow, magenta, and cyan signals with the black signal) There is a case that said to want to be high. In this case, as conventionally described in JP-A-7-203198 and the like, patterns are prepared for the number of controls, but the line buffer is large and the required patterns increase. The third embodiment has a mechanism for avoiding this problem.
[0022]
FIG. 3 is a block diagram of the third embodiment of the present invention, where 1 is an image input device such as a scanner (of course, image capture from a network may be used), and 2 is an input signal having a predetermined resolution (to the lower resolution side). Resolution conversion circuit (1) for conversion, 3 is an edge detection circuit for detecting a character edge region, 4-1 and 4-3 are line width detection circuits for determining whether or not the target portion is below the line width A, 4 -2 and 4-4 are line width detection circuits for determining whether or not the target portion is less than or equal to the line width B. 5-1 and 5-2 are line width A detection circuits 4-3 or line width B detection circuits 4. 4 is a resolution conversion circuit (2) for converting the determination result of -4 into a predetermined resolution (to the high resolution side). Reference numeral 8 denotes output signals (J, K, L) of the line width A detection circuit 4-1, the line width B detection circuit 4-2, the resolution conversion circuit (2) 5-1, 5-2, and the edge detection circuit 3. , M, N) is a comprehensive determination circuit that performs comprehensive determination.
[0023]
The differences from the first embodiment will be described below. In Embodiment 3, line width detection is performed on the original image data and on the image data after resolution conversion (50%). The line width is also detected with different widths (A and B), and the detection results (J, K, L, M) and edges of the circuits 4-1, 4-2, 5-1, 5-2 are prepared. A multistage signal is output by comprehensive determination with the output (N) of the detection circuit 3. That is, four types of line width information are obtained from the line width information of A and B and two different image resolutions. For example, if the detection is for a width of “3” (a value for relative comparison and the unit has no meaning), and if B is “4”, the resolution is halved. By doing so, it becomes possible to detect line widths such as “6” and “8”.
[0024]
The comprehensive determination circuit 8 outputs a determination signal (a signal for controlling the spatial filter and the UCR rate) according to the table and logical expression as shown in FIG. For example, application to applying a strong edge enhancement filter as the numerical value increases is possible.
[0025]
Although the embodiments of the present invention have been described above, the configurations of FIGS. 1 to 3 can be constructed using a computer system. In this case, the function of each circuit is realized by a so-called computer program. By loading this program into the computer, the intended functions of the circuits having the configurations shown in FIGS. 1 to 3 are achieved.
[0026]
【The invention's effect】
As described above, according to the present invention, the following effects can be obtained.
(1) In the character / line drawing area detection based on the line width information, low resolution data is used when detecting the line width information, so that even if the character width to be detected becomes large, it is possible without a large increase in hardware. Become.
(2) By preparing a plurality of image data with different resolutions, it is possible to prevent omission of detection of dense thin line character portions.
(3) In addition to the effects of (1) and (2), a signal for multistage control can be generated.
[Brief description of the drawings]
FIG. 1 is a block diagram of Embodiment 1 of the present invention.
FIG. 2 is a block diagram of Embodiment 2 of the present invention.
FIG. 3 is a block diagram of Embodiment 3 of the present invention.
FIG. 4 is an example of a dense pattern of lines.
FIG. 5 is an example of a table used for comprehensive determination according to the third embodiment.
FIG. 6 is a diagram for explaining conventional line width detection;
FIG. 7 is a diagram for explaining conventional multi-stage line width detection;
[Explanation of symbols]
1 Image Input Device 2 Resolution Conversion Circuit (1)
3 Edge detection circuit 4 Line width detection circuit 5 Resolution conversion circuit (2)
6 AND circuit 7 OR circuit 8 Comprehensive judgment circuit

Claims (2)

Inputting image data, the area detecting device for detecting a character-line region to the input image data,
A first means for detecting a contour of a character / line drawing of the input image data;
First, a region corresponding to a line width of two or more different characters / line drawings is detected for the input image data and at least one low-resolution image data obtained by converting the resolution of the input image data . Two means,
When the detection result of the first means indicates a contour, a value determined according to a predetermined table based on the detection result of the second means, and when the detection result of the first means indicates a non-contour A third means for outputting predetermined values as multi-stage determination signals ,
An area detection apparatus comprising: Inputting image data, the region detection method for detecting a character-line region to the input image data,
A first step of detecting a contour of a character / line drawing of the input image data;
First, a region corresponding to a line width of two or more different characters / line drawings is detected for the input image data and at least one low-resolution image data obtained by converting the resolution of the input image data . Two steps,
When the detection result of the first step indicates a contour, a value determined according to a predetermined table based on the detection result of the second step, and when the detection result of the first step indicates a non-contour A third step of outputting predetermined values as multi-stage determination signals ,
A region detection method characterized by comprising:

Images