JPH02274173A - Character separating device - Google Patents

Abstract

PURPOSE:To separate a character area on a dot with high accuracy by extracting an edge picture element from a smoothed picture and deciding it as a character area where the value of the count of an extracted edge picture element exceeds a prescribed value. CONSTITUTION:A smoothing processing section 1 applies smoothing filter to a digital multi-grade input picture (characters and dots are in existence in mixture such as a document picture or the like). A binarizing processing section 3 binarizes a differentiating processing picture by a differentiating processing section 2 at a specific level th1. A picture element made effective by the binarizing processing, that is, a picture element with a level over the level th1 or over is an edge picture element. A count processing section 4 counts the extracted edge picture element and decides the character area and the dot area depending on the count value P. Thus, not only characters on a white background but also characters on dots are separated with high accuracy.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a character separation device that extracts character areas on halftone dots from a digital multi-tone image.

[Conventional technology]

Document images often contain a mixture of halftone dots and characters.

A single character or line does not necessarily exist on a white background, but often a character or line exists on a halftone dot, as shown in the example shown in FIG.

When reproducing such document images with a digital copier 1
In order to improve image quality, it is preferable to remove moiré in the halftone dot area and to reproduce characters and lines clearly (without cutting them). In this case, it is necessary to separate and extract the characters within the broken line and the surrounding area several dots wide from the halftone dot area.

Conventionally, such separation methods are broadly classified into methods of extracting halftone dot areas using the properties of halftone dots and methods of extracting character areas using the properties of a single character.

The former example is described in Japanese Patent Laid-Open No. 60-80365, in which halftone dot pixels are detected by pattern matching, and a halftone dot area is defined within a mask of a size in which the halftone dot pixels exist.

An example of the latter is the paper "Moiré removal process using dither method"
(Sawara; Proceedings of the 2nd Non-Impact Printing Technology Symposium, 3-4.1985, pP, 69-7
2). In this case, in order to erase halftone dot edges, input data is first smoothed, edges of the smoothed data are emphasized, and then binarized, and portions with high threshold values are determined to be character areas.

[Problem to be solved by the invention]

However, according to the above-mentioned conventional technology, there is no problem in separating characters existing on a white background, but since the characters on the halftone dots are located near the halftone dots, the character area is buried in the halftone dots, resulting in poor accuracy. Cannot be separated well. In addition, in the method described in the above paper, it is difficult to set parameters to extract only the edges of a single character with high precision, so some of the edges of halftone dots also stand up, resulting in misidentification of halftone dot areas as character areas. There are things to do.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a character separation device for accurately separating character areas on halftone dots.

〔Example〕

FIG. 1 is a schematic block diagram showing one embodiment of the present invention. 1 is a smoothing processing section, 2 and 3 are differential processing sections and binarization processing sections for extracting edge pixels, and 4 is a counting processing section for counting and determining edge pixels.

The smoothing processing unit 1 applies, for example, a 3×3 smoothing filter to a digital multi-tone input image (such as a document image containing a mixture of characters and halftone dots). The value of k is selected in the range of about 1 to 10, and is determined depending on the character series of the target image, the number and area ratio of halftone dots, the characteristics of the scanner used to read the image, etc.

The differential processing unit 2 performs differential processing by applying, for example, a 3×3 Laplacian filter to the smoothed image.

The binarization processing unit 3 binarizes the differentially processed image by the differentiation processing unit 2 at a specific level thl. Pixels made valid by this binarization process, that is, pixels with a level of th1 or higher, are edge pixels.

In the counting processing section 4, the extracted edge pixels are counted in units of predetermined NXN blocks, and based on the counted value P, a character area and a halftone dot area are determined.

That is, in the case of processing one pixel at a time, when the count value P exceeds a certain value th2, the center pixel of the block of interest is set as a character area, otherwise it is set as a halftone dot area.

Alternatively, it is also possible to process in blocks,
In this case, the pixels of the entire block where the count value P is equal to or greater than th2 are defined as a character area, and the pixels of the entire block where this is not the case are defined as a halftone dot area.

According to such processing, even if some of the halftone edge pixels are erroneously extracted in the previous processing, the halftone edge pixels can be removed because they have a low density. Furthermore, even if a part of a character is faded, if the area around it is extracted as an edge, the faded part can be determined as a character area.

Note that N, thl, and th2 are determined to appropriate values through experiments for parameter setting.

FIG. 2 is a flowchart of the algorithm for the above-mentioned character separation process. Note that the filter for smoothing and differentiation is a 3x3 filter.

Three lines of image data are set at the beginning of processing.

FIG. 3 shows an example of both signal processing systems of a digital copier to which the present invention is applied. If the reproduced image is monochrome, the image signal input to this image signal processing system is the luminance signal of the document, but this may also be a luminance signal generated from RGB signals. If the reproduced image is in color, the RGB signals of the color-separated and read document or the YMC signals after color correction are input, and each color signal is processed in the same way as the luminance signal.

In FIG. 3, the input image signal is accumulated in a buffer 12 and smoothed by a smoothing section 12. This smoothing section 12 serves both as the smoothing processing section 1 for character separation in FIG. 1 and as a preprocessing section for preventing the occurrence of moiré due to dithering of halftone dots.

The smoothed image signal is accumulated in the buffer 13 and is input to the halftone/character area determining section 14. This halftone/character area determination section 14 is a processing section corresponding to the first differential processing section 2 , the binarization processing section 3 , and the counting processing section 4 , and its determination output signal becomes a control signal for the multiplexer 15 .

The halftone area processing section 16 is a section that performs dither processing on the smoothed image signal inputted from the buffer 13.

Here, we emphasize gradation and use a spiral dither matrix. An example of such a dither matrix is shown in FIG.

The character area processing section 17 performs sharpening processing on the image signal input from the buffer 11 (input image signal from the outside), and then performs dither processing. The sharpening process is performed by applying a filter such as the one shown below. The dither processing uses a Bayer type dither matrix that emphasizes resolution. An example thereof is shown in FIG.

The multiplexer 15 outputs an image signal inputted from the halftone area processing section 16 when the judgment output of the halftone/character area judgment section 14 indicates a halftone dot area, and outputs an image signal inputted from the halftone dot area processing section 16 when the judgment output of the halftone dot/character area judgment section 14 indicates a text area. The image signal input from the character area processing section 17 is used as an output image signal.

Here, the character area is a character and its range several dots wide (inside the broken line in FIG. 6).

In this way, it is possible to remove moiré in the halftone dot area and reproduce an image in which the characters and lines in the character area are clear and unbroken. In the case of color, similar processing is performed for each color signal, so the quality of the reproduced image can be improved.

〔Effect of the invention〕

As explained above, according to the present invention, not only characters on a white background but also characters on halftone dots can be separated from the halftone dots with high precision.
The image quality of the reproduced image can be improved by applying optimal processing to the text area and halftone area.

[Brief explanation of drawings]

Fig. 1 is a schematic block diagram of an example of a character separation device of the present invention, Fig. 2 is a flowchart showing an algorithm for character separation processing, and Fig. 3 is a schematic block diagram of an image signal processing system of a digital copier to which the present invention is applied. 4 and 5 are diagrams showing examples of dither matrices, and FIG. 6 is an explanatory diagram of a character area. 1... Smoothing processing section, 2... Differential processing section, 3...
・Binarization processing section, 4... Counting processing section, 2... Smoothing section, 11.13... Buffer, 4... Halftone dot/
Character area determination section, 5...Multiplexer, 6. Halftone area processing section, 7. Character area processing section. Book... ・―― o＃h

Claims (1)

[Claims] (1) A means for smoothing a digital multi-tone image, a means for extracting edge pixels from the smoothed image by the means, and a means for counting the edge pixels extracted by the means, and the count value exceeding a predetermined value. 1. A character separation device for extracting a character area on a halftone dot, comprising means for determining a character area on a halftone dot.