JPS63114462A - Identifying device for image area - Google Patents

Abstract

PURPOSE:To identify the image areas of a character, a line drawing, a photograph and a dotted photograph in image data and judging the picture hue of a noticed picture element based on results obtained by comparing the difference between the average density value of a block and the density value of the noticed picture element, the low and the high frequency components of a special frequency in the block with first, second and third thresholds, respectively. CONSTITUTION:Image data whose gradation is corrected by a gradation correction circuit 3 is inputted to a block making circuit 13 and a picture element in the prescribed size is segmented. A density difference detection circuit 7 takes the absolute value of the difference between the average density value and the density value of the noticed picture element. If the absolute value is smaller than a first threshold, the noticed picture element is decided to be in a halftone image area or in a background area, otherwise, it it decided to be in the image area of which density partly and violently changes. A space filter circuit 8 detects the image edge of a comparatively low frequency and compares it with a second threshold. If the edge is bigger, it is decided that a low frequency edge exists. A space filter 9 compares the image edge with a third threshold, and if the edge is bigger, it is decided that there is a high frequency edge including the dot. Based on these results, a decision circuit 10 identifies the image data on an original in which the photograph area, the character area and the dotted photograph area mixedly exist.

Description

Detailed Description of the Invention [Industrial Application Field] The present invention relates to an image forming apparatus such as a digital copying machine, which uses an image area to identify the image tone (characteristics and properties of an image) of a certain image area in a document. This relates to identification devices.

[Prior Art] Conventionally, this type of image forming apparatus uses C as shown in FIG.
The data read by the input sensor 21 such as OD is input to A at 22.
/D conversion, then gradation correction such as γ conversion is performed at step 23, and the signal is converted into a dot ON10FF signal at binarization circuit 24, which is output by a binary printer 25.

Typical binarization methods in the binarization circuit 24 include a method of binarization using a fixed threshold value and a method of binarization using a dither method.

[Problems to be Solved by the Invention] With the above fixed threshold binarization method, the reproduced image quality of characters and line drawings in the manuscript is good, but in the case of photographs, gradation is lost and the image quality is poor. There was a problem with the decline. On the other hand, with the dither method, on the other hand, the gradation of the photo is good, but text and line drawings are cut off, and moiré fringes occur in halftone photos with a certain number of screen strokes, resulting in a decrease in image quality. There was a problem. For this reason, when characters, line drawings, photographs, and halftone photographs coexist in the same document, it is difficult to form all images satisfactorily.

SUMMARY OF THE INVENTION Accordingly, the present invention proposes an image area identification device that eliminates the above-mentioned drawbacks of the prior art and faithfully identifies the image area of image data of a document containing a mixture of characters, line drawings, photographs, and halftone photographs.

[Means for Solving the Problems] The configuration of the image area recognition device of the present invention proposed to achieve the above-mentioned problems is to divide image data into blocks and recognize image areas in the blocks. The recognition device includes a detection unit that inputs the block and detects a difference between an average density value of the block and a density value of the pixel of interest, and a detection unit that compares the density difference with a first threshold value. a first comparison means for inputting the block and extracting a low frequency component of a spatial frequency within the block; and a second comparison means for comparing the low frequency component with a second threshold value. means, a second filter means for inputting the block and extracting a high-frequency component of a spatial frequency within the block, a third comparing means for comparing the high-frequency component with a third threshold; and determining means for determining the image tone of the pixel of interest within the block based on the two comparison results.

[Operation] In the above configuration, the comparison result by the first comparison means indicates whether it is a halftone image such as a photograph, the result of the second comparison means indicates whether it is a line drawing area, and the comparison result of the third comparison means indicates whether it is a halftone image such as a photograph. The result indicates whether it is a halftone area.

[Examples] Examples of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an embodiment in which the image area recognition device of the present invention is applied to a printing device that reads an image, processes the image, and then prints the image. According to this printing apparatus, a document (not shown) is read by an input sensor 1 such as CC or D, and is converted into, for example, an 8-pit digital signal by an A/D converter 2. This digital signal represents 256 gradations of light and shade. The gradation correction circuit 3 performs various gradation operations on this signal.

Next, this signal enters three types of binarization circuits: a fixed threshold binarization circuit 4, a dithering binarization circuit 5, and a moiré suppression binarization circuit 6. The fixed threshold binarization circuit 4 compares a certain threshold value with the image data, and when the image data is larger, it is set as "1°".
°, and conversely, when it is small, it is binarized by outputting 0°°. The dithering circuit 5 compares the image data with the image data in the order of the threshold values indicated by the dither matrix, and similarly outputs a signal of °1" or O". This digitalization is effective in binarizing halftone (continuous tone) images. The moire suppression binarization circuit 6 performs binarization in a manner that does not generate moire fringes even when the original is a halftone photograph. As a method of suppressing such moire, for example, a method of adding smoothing processing to the image data to remove periodic components in the image data and reducing the moire by t[1], or a method that reduces the number of dots in the original and makes beats less likely to occur. There are methods such as selecting a dither matrix size, or randomly selecting dither matrix components to suppress beats.

The outputs of the three binarization circuits are appropriately switched by the switch 11 according to the judgment result of the judgment circuit 10, and sent to the printer 12 as a printer dot 0N10FF signal to form an image.

The operation of the determination circuit 10 and the like will be explained. In synchronization with the input of the image data whose tone has been corrected by the tone correction circuit 3 into the three binarization circuits, the same image data is input into the blocking circuit 13. This blocking circuit 13 cuts out a pixel block of a predetermined size (for example, 5×5 size as shown in FIG. It is sent to a spatial filter circuit 8 and a second spatial filter circuit 9. The density difference detection circuit 7 takes the absolute value of the difference between the average value within the specific pixel block and the pixel of interest. Generally, when the absolute value of this difference is smaller than a certain threshold, the pixel of interest is in a halftone image area such as a photographic image.
Or it can be said that it is within the area of the background image. On the other hand, when this absolute value is large, it can be said that the image area has a large local density change, such as a character portion or a halftone image. Therefore, when the density difference detection circuit 7 determines that the pixel of interest is in the halftone image area, it outputs, for example, °゛O゛, and when it determines that the pixel of interest is in the character part, point tA image area, it outputs 1''. Make it.

Next, the first spatial filter circuit 8 filters a relatively low frequency (
A spatial filter is configured to perform edge detection at a rate of approximately 2 lines up/mm). For example, this is shown in Figure 3 (a) and (b).
This can be realized by combining first-order differential filters with coefficients as shown in .

The second spatial filter circuit constitutes a spatial filter that detects edges of halftone dot frequencies.

This is, for example, 2 with the coefficients shown in Figure 3(C).
This can be realized using an order differential filter. In general, considering the characteristics of a filter, if the kernel size is the same, a negative-order differential filter has a peak of transmittance on the lower frequency side than a second-order differential filter. Figure 4(a) shows 2
This is a one-dimensional model showing the characteristics of the fffl spatial filter. -order differential filter is, for example, approximately 2 degrees p/
The second-order differential filter is selected to have a characteristic that has a peak at about 4 n p /mm, for example.

On the other hand, FIG. 4(b) shows the spatial frequency characteristics of the text portion and halftone photograph portion in the original image. In terms of frequency, the power spectrum of characters decreases almost continuously from the low frequency side to the high frequency side, whereas in halftone photographs, the power spectrum decreases discontinuously and periodically at the high frequency component corresponding to the halftone frequency. The difference is that there is a peak in the power spectrum.

Therefore, in the first spatial filter circuit 8, since the transmission band is at 21 p/mm, the spatial frequency components of the characters are detected, but the spatial frequency components of the halftone photograph are not detected.

On the other hand, the second spatial filter circuit 9 has a transmission band of 4 p
/ mm and also has a wide half-value width of the peak that indicates the filter characteristics, so this filter detects the frequency components of characters centering on the spatial frequency of the halftone photograph.

FIG. 5 is a detailed block diagram of circuit 10 of FIG. 1 for determining image regions within an image. In the block diagram of FIG. 5, a line memory 50 is used as the blocking circuit 13. If a 5×5 matrix as shown in FIG. 3 is used, the line memory 50 only needs to be prepared for five lines. The concentration difference detection circuit 7 includes an average value calculation circuit 51 and a subtracter 52. The average value calculation circuit 51 receives one block of pixel data from the line memory 50, performs addition and division, and calculates the average value within the block. The density of the pixel of interest (in the case of a 5×5 matrix, the center pixel) in the line memory 50 is input to one input of the subtracter 52. Then, the subtracter 52 takes the absolute value of the difference between the average value and the pixel density of interest.

The absolute value of this difference is compared in a comparator 57 with a specific threshold T1. When the difference between the image data of the pixel of interest and the average value is small, 8 is determined as a half-tone photographic area or a background area of a character image, and "0" is output. Conversely, when the difference is large, it is determined that the area is a character area or a halftone photograph area, and "1" is output. This output result A is input to the determination circuit 10.

Further, the image data of the pixels in the block from the line memory 50 is inputted to the first spatial filter circuit 8, that is, inputted to the first-order differentiation circuits 53 and 54 in FIG.
A first-order differential operation as shown in (b) is performed in each differentiator circuit, and the absolute value is output. The differentiation circuit includes a multiplier that multiplies the pixel density by a matrix coefficient corresponding to that pixel,
It consists of an adder that calculates the sum of their products. As a result, the circuit 53 detects relatively low frequency image edges in the vertical direction, and the circuit 54 detects low frequency edges in the horizontal direction. These outputs are summed in an adder 56 and then compared with a specific threshold T2 in a comparator 58. The addition result is the threshold T
When it is larger than 2, it is determined that there is a low frequency edge and the angle is 1°.
° is output, and conversely when it is small, it is “0” as there is no edge.
Output. This output result B is input to the determiner 10 in the same way as A.

In the second spatial filter circuit 9, a second-order differential circuit 55 inputs the pixel data in the block, performs a second-order differential operation as shown in FIG. 3(C), and outputs the absolute value. This value is compared in a comparator 59 with a specific threshold T3. When the second-order differential value is larger than the threshold T3, it is determined that there is a high-frequency edge including halftone dots, and 1'' is output; on the other hand, when it is smaller,
Outputs “0” as there is no edge. This result C is input to the determiner 10.

Table 1 Here, X may be 0" or 1".

The determining device 10 determines the photo area and character area based on the three output results A, B, and C as shown in Table 1. When output A is "1" and outputs B and C are both "1", it is determined that the area is a character area. Also, when B is 0''° and C is ``1'', it is determined to be a halftone area.Other combinations of the outputs of A, B, and C are possible, but the probability of them actually occurring is very low, so ignore them. .

The determiner 10 can be constructed by a simple circuit as shown in FIG. The output A passed through the inverter 70 is directly used as a photo area designation signal, and the outputs of AND gates 71 and 72 are used as a character area designation signal and a halftone dot region designation signal, respectively. These instruction signals cause the switch 11 shown in FIG. 1 to be switched. That is, the binary image data output from the switch 11 becomes image data subjected to IA processing by binary conversion most suitable for the image area, so that the image reproduced by the printer 12 is of high quality.

[Effects of the Invention] As explained above, according to the image area recognition device of the present invention,
Image data of an original image containing a mixture of photo areas, character areas, and halftone photo areas can be identified according to these areas.

[Brief explanation of the drawing]

Fig. 1 is a block diagram schematically showing an embodiment in which the image area recognition device of the present invention is applied to a printed lid, Fig. 2 is an explanatory diagram of a conventional binary image forming device, and Fig. 3 (a) is a matrix diagram representing the coefficient values of the first-order differential filter for vertical edge detection, FIG. 3(b) is a matrix diagram representing the coefficient values of the first-order differential filter for horizontal edge detection, and FIG.
C) is a matrix diagram showing the coefficient values of the second-order differential filter for detecting high-frequency edges; FIG. Figure (b) is a characteristic diagram showing the spatial frequency characteristics of typical character areas and halftone dot areas. Figure 5 is a detailed block diagram of a circuit for image area determination. Figure 6 is a configuration example of a determination circuit. FIG. In the figure, 4...Fixed threshold value binarization circuit, 5...Dizzani value conversion circuit, 6...Moiré suppression binarization circuit, 7...Density difference detection circuit, 8...First Spatial filter circuit, 9... Second spatial filter circuit, 10... Judgment circuit, 11...
- Switch, 12... Printer. Patent applicant Canon Co., Ltd. Nada β (a) l ρp/mm (b) Figure 4 Figure 6

Claims (3)

[Claims] (1) In an image area recognition device that divides image data into blocks and recognizes the image area of the pixel of interest in the block, the block is input and the average density value of the block and the density value of the pixel of interest are calculated. a detection means for detecting a difference; a first comparison means for comparing the magnitude of the density difference with a first threshold value; and a first comparison means for inputting the block and extracting a low frequency component of a spatial frequency within the block. a second comparing means for comparing the low frequency component with a second threshold; and a second filter means for inputting the block and extracting the high frequency component of the spatial frequency within the block. an image area comprising: a third comparing means for comparing the high-frequency component with a third threshold; and a determining means for determining the image tone of the pixel of interest in the block based on the three comparison results. Identification device. (2) The image area identification device according to claim 1, wherein the first filter means is a first-order differential filter. (3) The image area identification device according to claim 1, wherein the second filter means is a second-order differential filter.