JPH09259273A - Image similarity decision device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve decision precision without increasing a circuit scale by deciding similarity by matching the number of pixels of each color and features of an object image against data in a specific image data memory. SOLUTION: Comparing circuits 31-36 of a decision circuit include multipliers 44 and 46 which multiply specific image data inputted from the specific image data memory 30 by values from registers 40 and 42 where permissible ranges from the data are set, and comparators 48 and 50 compare label image data inputted from a label image data memory 26 to check which of the two permissible rages the data belongs to, and outputs comparison results F1-F6 through an AND circuit 52. Then a final decision circuit 54 counts the comparison results F1-F6 outputted from the respective comparing circuits 31-36 by a counter 56, whose counting result is compared with threshold values set in registers 58 and 60 to output the two decision results. Thus, the similarity between the image data and a specific image is decided according to the number of pixels and features.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image similarity determination device for determining the similarity between image data and a specific image, and more particularly to a scanning type color image reading device such as a color image scanner or a color copy. The present invention relates to an image similarity determination device which is suitable for, and is capable of highly accurately determining the similarity between a specific image that should be prohibited from being captured.

[0002]

2. Description of the Related Art With the recent increase in image quality of full-color image forming apparatuses, there is an increasing risk that a manuscript having a character that should not be copied, such as banknotes, revenue stamps, and stamps, will be copied. In order to solve such a problem, a technique has been proposed in which the content of a document is determined and the image formation is prohibited for images that should not be copied.

For example, in Japanese Unexamined Patent Publication No. 5-14685, different R, G and B determination circuits are provided for each specific image and each include three window comparators for each color.
A value is recorded, and the stored R, G, and B values are compared with the R, G, and B values of the color original image, a pattern corresponding to the comparison result, that is, the certainty factor is generated and output, and then the color original is printed. It is described that an output image based on the R, G, and B values of the image is output, and the pattern and the image in which the color original image is reproduced are combined in the same output image.

Further, in Japanese Patent Laid-Open No. 6-113128, input data is thinned out, and whether or not the tint matches the specific original in pixel units is passed through a look-up table, and a preset register value is set by an integral value in 3 line units. In contrast to the four-stage judgment, the tint for each specific original is judged, while the maximum value of the judgment signal from the start of copying to the present is held and output is performed according to the level of possibility that the specific original is included It is described to modulate.

[0005]

However, in any of the techniques, since the input image is directly compared with the specific image, not only the determination takes time but also the determination accuracy is low.

Particularly, in Japanese Laid-Open Patent Publication No. 5-14685, it is necessary to provide a determination circuit including three window comparators for each specific document, so that the circuit becomes large in scale in proportion to the number of specific images. Further, even in the basic configuration, a plurality of line memories are required and the circuit is large. Further, the window comparator needs to be large to some extent, so that the circuit is also large. Further, when the original is skewed, there is a problem that the effectiveness is doubtful.

Further, in Japanese Patent Laid-Open No. 6-113128, since the determination is made in units of three lines, it is necessary to determine the similarity in the middle of a specific image. Therefore, the modulation is performed according to the level of the similarity. Not only is such a troublesome process required, but the determination is unfavorable. Further, there is a problem in that it is necessary to increase the tint determination circuit when the types of specific images are increased.

The present invention has been made to solve the above-mentioned conventional problems, and it is an object of the present invention to improve the determination accuracy without increasing the circuit scale.

[0009]

SUMMARY OF THE INVENTION According to the present invention, in an image similarity determining apparatus for determining the similarity between image data and a specific image, the specific image storing the number of pixels and the feature of each color of each specific image is stored. A data memory, a means for scanning the image data, labeling and extracting the target image, and a means for rescanning the image data and obtaining the number of pixels and characteristics of each color for each target image extracted in the first assessment. And a means for comparing the number of pixels of each color of the target image with the data of the specific image data memory to determine the degree of similarity, thereby solving the above problem.

Further, the first scanning of the image data is performed in black and white and with the focus defocused.

In the present invention, the image data is scanned, for example, in black and white and the focus is defocused, the target image is labeled and extracted by a known labeling technique, and the area to be the target image is specified. Then, the image data is rescanned in color, and the number of pixels (area) and features (color components) for each color component such as RGB, Y, Cb, Cr, or CMYK is rescanned for each target image extracted in the first scan. To get And
The number of pixels of each color and the characteristics of the target image are compared with the data of the specific image stored in advance to determine the degree of similarity. Therefore, the similarity can be determined without increasing the circuit scale, only by increasing the number of memories of the specific image data memory corresponding to the number of specific images. Further, since the area to be the target image is specified by the first scanning, it is possible to quickly make a highly accurate determination.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a flow chart showing the processing procedure of this embodiment.

An input image 10 to be read from a color scanner, a copying machine or the like is as shown in FIG. 2, for example, specific images a, b, c of stamps, banknotes, securities, etc.
It is assumed that d and e are included. First black-and-white scanning is performed on such an image reading surface (step 100) to obtain a binarized image 12 as shown in FIG. 3, and target graphics are a ′, b ′, c ′, d. Label as ', e' (step 110). At this time, the focus of scanning of the target image is defocused so that the target image is not divided and the concatenated relationship is maintained. However, it does not extend to the adjacent target image. At this point, the binarized black-and-white data is the target (not necessarily the color).

Next, from the image data labeled as shown in FIG. 3, as shown in FIG. 4, an end point (for example, an upper left point (X
1, Y1) and the lower right point (X2, Y2)) are calculated (step 120). The calculation result is as shown in FIG. 5, for example.

Next, a second color scan is performed (step 130), and as shown in FIG. 5, the upper left point (X
1, Y1) and the lower right point (X2, Y2), the area and component of each color are calculated for each area including the target image (step 140). The calculation here may be a simple addition. Moreover, each color means RGB, Y, Cb, Cr, C.
Refers to MYK and the like.

After color scanning, label image memory data as shown in FIG. 6 is generated for each of the obtained target images (step 150).

When the label image data is generated in step 150, the variable sampling circuit 20 can be provided on the input side as shown in FIG. 7 to thin out the input image data according to its density. That is, if high-density loading is performed, thinning-out is increased, and if low-density loading is performed, thinning-out is reduced to prevent an increase in the label image data memory capacity. Then, the label image data is read from the label image data memory 26 using the label value for the input image data as an address. The area is added to the read value by the adder 22, the chromaticity (component) is integrated by the integrator 24, and the result is written in the same address.
By repeating this until the input image data is completed, a label image memory having area and chromaticity data corresponding to R, G, and B for each label number (memory address) as shown in FIG. Data is generated.

Then, it is collated with a specific image data table as shown in FIG. 8 (step 160). In this specific image data table, the area and chromaticity of each color, for example, R, G, B are stored. At this time, it is judged whether or not it is within the range of the reference color (step 170),
Those within a certain range are determined to have a high angle of similarity (step 180).

At the time of the determination in steps 170 and 180, the data of 6 items for each label image as shown in FIG. 6 and the 6 items of the specific image data as shown in FIG. For example, the comparison is performed using a determination circuit as shown in FIG. The determination circuit is provided with comparison circuits 31 to 36 corresponding to each item. Each comparison circuit includes multipliers 44 and 46 for multiplying the specific image data input from the specific image data memory 30 and the values from the registers 40 and 42 for setting the allowable range from the specific image data, and the label image data memory 26. The comparators 48 and 50 compare which of the two allowable ranges the label image data input from the
Output through the ND circuit 52. The final determination circuit 54 is
Comparison results F1 to F output from the comparison circuits 31 to 36
6 is counted by the counter 56, and the counting result is registered in the register 58,
Comparing the threshold value set to 60 with the comparators 62 and 64,
Two determination results 1 and 2 are output. That is, the comparison for each item is set to "1" if it is within a certain plus or minus range set by the registers 40 and 42, and is set to "0" otherwise. Six values F obtained by this comparison
A final determination of similarity is made by 1 to F6. Of the six "1" s or "0" s, the number of "1" s or more determines whether the determination result is "1" is determined by the values of the registers 58 and 60. In this determination, the similarity determination can be output in three stages by, for example, comparing two ways from the two registers 58 and 60. That is, if the two bits of the determination results 1 and 2 of the determination result output are both “0”, it is determined that they are not similar, and the two register values,
For example, when the register 58 is "3" and the register 60 is "5", if the determination result output determination result 1 is "1", it is determined that they are similar, and the determination result 2 is "1". If the result 1 is “0”, it is determined that there is a possibility of similarity.

When an image having a high degree of accuracy similar to the specific image is input, the scheduled image reading is stopped or the read data is processed so that the input image is not reproduced as it is (step 190).

In this embodiment, since the first scan is performed in black and white, the process is simple. It is also possible to perform the first scanning in color to improve the labeling accuracy.

Further, in this embodiment, since the area of each color is obtained by adding the number of pixels of each color and the characteristic is obtained by integrating the chromaticity, the processing is simple. The method of determining the area and the component is not limited to this.

[0024]

According to the present invention, the degree of similarity between image data and a specific image can be accurately determined based on the number of pixels and characteristics. Further, the addition or change of the specific image can be easily performed by simply adding or changing the data in the specific image data table.

[Brief description of drawings]

FIG. 1 is a flowchart showing a processing procedure in an embodiment of the present invention.

FIG. 2 is a diagram showing an example of an input image processed by the embodiment.

FIG. 3 is a diagram showing an image that has been binarized and labeled after black-and-white scanning.

FIG. 4 is a diagram showing end point coordinates of a labeled target image.

5 is a diagram showing a table in which end point coordinates are extracted from FIG.

FIG. 6 is a diagram showing an example of label image memory data generated in the embodiment.

FIG. 7 is a block diagram showing a configuration example of a label image data generation circuit used in the embodiment.

FIG. 8 is a diagram showing an example of a specific image data table used in the embodiment.

FIG. 9 is a block diagram showing a configuration example of a determination circuit in the same manner.

[Explanation of symbols]

 10 ... Input image a, b, c, d, e ... Specific image a ', b', c ', d', e '... Labeling image 12 ... Binary image 20 ... Variable sampling circuit 22 ... Adder 24 ... Integrator 26 ... Label image data memory 30 ... Specific image data memory 31-36 ... Comparison circuit 40, 42, 58, 60 ... Register 44, 46 ... Multiplier 48, 50, 62, 64 ... Comparator 52 ... AND circuit 54 … Final judgment circuit 56… Counter

Claims (2)

[Claims] 1. An image similarity determination device for determining the similarity between image data and a specific image, wherein a specific image data memory for storing the number of pixels and features of each color of each specific image, and image data scanning. Then, a means for labeling and extracting the target image, a means for rescanning the image data, a means for obtaining the number of pixels and features of each color for each target image extracted in the first assessment, and a pixel for each color of the target image An image similarity determination device comprising: means for comparing the number and characteristics with the data in the specific image data memory to determine the similarity. 2. The image similarity determination device according to claim 1, wherein the first scan of the image data is performed in black and white and with a defocused focus.