JP2000076458A - Color picture recognizing device, its method, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To precisely detect a specific symbol from an inputted picture with a simple constitution. SOLUTION: A specific color detecting circuit 21 extracts the picture of the constitution color of a specific symbol from an inputted picture, an intra- circular area on pixel total number detecting circuit 25, an intra-circular area on/off inversion number detecting circuit 26, a first on-circumference on pixel total number detecting circuit 27 and a first on-circumference on/off inversion total number detecting circuit 28 detect a characteristic value concerning the specific symbol from a picture within an area while scanning the picture extracted by the circuit 21 in the area of a size according to the specific symbol. A second on-circumference data coincidence calculating circuit 40 calculates a coincidence degree based on a picture pattern expressing the picture positioned on a second circumference with the center pixel of the area as its center and a previously set normal pattern among the picture extracted by the circuit 21 and a judging circuit 29 judges the presence/absence of the specific symbol by using the detecting result and the calculation result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color image recognition technique for detecting a specific symbol from input image data, and more particularly to an apparatus and a duplication for re-inputting an already output object and improving the image quality of an output object when outputting. The present invention relates to a color image recognition device and a color image recognition method suitable for use in a device having a security function, and a recording medium storing a computer program for realizing the color image recognition method.

[0002]

2. Description of the Related Art In recent years, colorization of output devices such as copying machines and printers has been advanced. Further, at present, as a result of rapid progress in color image recording apparatuses using digital data processing techniques, it is possible to obtain output images of sufficiently high quality. For this reason, many copying machines have been put to practical use in which a special manuscript whose copying is prohibited by law is determined and illegal copying is prohibited. Usually, the reference of a special document is the presence or absence of a symbol that always exists in the special document, and a pattern matching method is generally used as a method for detecting the symbol. Conventional copiers that employ a pattern matching method to prohibit illegal copying first include the angle of a document placed on the copier (for example, 0 degrees, 90 degrees,
180 degrees, 270 degrees), the reference pattern is corrected at the detected angle, the presence or absence of the symbol is determined by the pattern matching method, and it is determined whether or not the original is a special original. / Determine ban.

On the other hand, in offices of companies and the like, in addition to the above-mentioned special manuscripts, there are a large number of documents whose copying is prohibited from the viewpoint of importance of contents and confidentiality. Generally, such a copy-prohibited document is stamped by a manager of the document so as to be distinguished from other copyable general documents so as to prohibit copying. The owner of the document confirms whether or not the seal is stamped on the document, and can freely copy the seal as necessary if the seal is not stamped.

[0004]

However, in the method in which the administrator stamps a predetermined seal on a copy-inhibited document to notify that the copy is prohibited, it is impossible to reliably prohibit the copy of the copy-inhibited document. Can not. For example, it is conceivable that the holder of the copy-inhibited document overlooks the seal and copies the copy-inhibited document, or inputs and saves it in the image filing device. That is, the above-described method is a management method that expects the attention and good intentions of the holder of the copy-forbidden document, and cannot reliably exclude the copy of the copy-forbidden document. As is apparent from this, in order to reliably eliminate the duplication of the copy-inhibited document, no human should be involved in the process of determining the copy-inhibited document.

As a method of excluding a person from the process of determining a copy-inhibited document, it is conceivable to simply use a conventional copying machine that prohibits copying of a special document. In this case, the accuracy of the determination can be reduced with a small amount of hardware resources. It is difficult to keep it high enough. As described above, the seal on the copy-inhibited document is generally stamped by the administrator (that is, a person) of the document. It is extremely difficult to exactly match the orientation of the copy-inhibited document (the direction of the character string or the paper) with the direction of the seal by the administrator, and it is not required. Therefore, the orientation of the seal and the copy-inhibited document are not required. It usually does not match the direction of. Therefore, even if the direction of the original placed on the copier is detected, the determination of the copy-inhibited document (detection of the seal) fails very often.

[0006] Also, reference patterns are created and stored by the number of assumed rotation angles of the seal, and pattern matching is performed using all of these reference patterns, or an input image is formed at all assumed rotation angles of the seal. Is rotated, and the pattern matching is performed using one reference pattern for each input image after the rotation, the accuracy of the determination can be maintained sufficiently high. In addition to this, a large-capacity storage device that can be accessed at high speed is required. Furthermore, due to the nature of the pattern matching method, it is necessary to process the image with a higher resolution in order to improve the accuracy of the determination,
The level of required hardware resources also increases.

The present invention has been made in view of the above circumstances, and detects a specific symbol from an input image with high accuracy using less hardware resources as compared with a case where a general pattern matching method is applied. It is an object of the present invention to provide a color image recognition device and a color image recognition method capable of performing the method, and a recording medium storing a computer program for realizing the color image recognition method.

[0008]

According to a first aspect of the present invention, there is provided a color image recognizing apparatus which receives a color image signal and converts a specific symbol from an image represented by the color image signal. In a color image recognition apparatus for recognizing, a specific color extracting unit for extracting an image of a constituent color of a specific symbol from the image, and scanning the image extracted by the specific color extracting unit in an area of a size corresponding to the specific symbol. A characteristic value detecting means for detecting a characteristic value relating to the specific symbol from an image in the area; and an image extracted by the specific color extracting means, which is located on a circumference centered on a central pixel of the area. A coincidence calculating means for calculating a degree of coincidence between the image in the area and the specific symbol based on an image pattern representing the image and a predetermined regular pattern; The presence or absence of the specific symbol is determined based on the characteristic value detected by the means, the coincidence calculated by the coincidence calculating means, and an allowable range set in advance for the characteristic value and the coincidence. And a determination means for outputting a determination result as a recognition result. According to this configuration, the image data for performing the pattern comparison is only the image data on the circumference, and the hardware resources required for the comparison process can be reduced as compared with the case where all the image data in the area are compared. . In addition, since the circumference around the center pixel of the area is used, even if the input image is rotated, it is possible to cope with it by shifting the start position of the regular pattern. If the circumference is set at an appropriate position, the probability of occurrence of erroneous recognition can be sufficiently reduced. Furthermore, since the determination is made in consideration of not only the degree of coincidence of the pattern but also the characteristic value, a higher recognition rate can be achieved by adopting a characteristic value that does not depend on the rotation of the input image.

Further, in the above configuration, the content of the calculation performed by the coincidence calculating means on a value of a block corresponding to each pixel of the regular pattern based on a state of each pixel of the image located on the circumference. May be determined, and the degree of coincidence may be calculated based on the result of the operation of the operation content. Alternatively, an indicator may be provided for indicating the diameter of the circumference, and the regular pattern may be a pattern corresponding to the diameter indicated by the indicator. Alternatively, the area may be a circular area or an annular area, or one symbol may be selected from a plurality of symbols, and the selected symbol may be recognized.

According to another aspect of the present invention, there is provided a color image recognition apparatus which receives a color image signal and recognizes a specific symbol from an image represented by the color image signal. In the recognition device, a specific color extracting unit that extracts an image of a component color of a specific symbol from the image, and among the images extracted by the specific color extracting unit,
A match between the image in the area and the specific symbol based on an image pattern representing an image located on the circumference centered on the center pixel of the area of the size corresponding to the specific symbol and a preset regular pattern Determining the presence or absence of the specific symbol based on the degree of coincidence calculated by the degree of coincidence calculated by the degree of coincidence calculation and an allowable range preset for the degree of coincidence. Determining means for outputting as a recognition result.

According to this configuration, the image data for pattern comparison is only the image data on the circumference, and the hardware resources required for the comparison processing are reduced as compared with the case where all the image data in the area are compared. be able to. Moreover,
Since the circumference around the center pixel of the area is used, even if the input image is rotated, it can be handled by shifting the start position of the regular pattern. If the circumference is set at an appropriate position, the probability of occurrence of erroneous recognition can be sufficiently reduced.

According to another aspect of the present invention, there is provided a color image recognition method comprising the steps of: inputting a color image signal; and recognizing a specific symbol from an image represented by the color image signal. In the recognition method, a specific color extraction step of extracting an image of a constituent color of a specific symbol from the image, and scanning the image extracted in the specific color extraction step in an area of a size corresponding to the specific symbol,
A characteristic value detecting step of detecting a characteristic value relating to the specific symbol from an image in the area; and an image located on a circumference centered on a central pixel of the area among the images extracted in the specific color extracting step. A matching value calculating step of calculating a matching degree between the image in the region and the specific symbol based on an image pattern representing a predetermined regular pattern, and a characteristic value detected in the characteristic value detecting step, The degree of coincidence calculated by the degree of coincidence calculation means,
Determine the presence or absence of the specific symbol based on the characteristic value and the tolerance set in advance for the degree of coincidence,
A determination step of outputting the determination result as a recognition result.

The color image recognition method according to claim 9, further comprising the steps of: inputting a color image signal; and recognizing a specific symbol from an image represented by the color image signal. A specific color extracting step of extracting an image of the constituent color of the image, and an image located on a circumference centered on a central pixel of an area of a size corresponding to the specific symbol among the images extracted by the specific color extracting step A matching degree calculating step of calculating a matching degree between the image in the region and the specific symbol based on an image pattern representing the image pattern and a preset regular pattern; and a matching degree calculated in the matching degree calculating step, A determination step of determining the presence or absence of the specific symbol based on a predetermined allowable range for the degree of matching, and outputting a determination result as a recognition result. It is characterized in that.

According to each of the above-described methods, the image data for pattern comparison is only the image data on the circumference, and the hardware resources required for the comparison processing are smaller than when all the image data in the area are compared. Can be reduced. In addition, since the circumference around the center pixel of the area is used, even if the input image is rotated, it is possible to cope with it by shifting the start position of the regular pattern. Also,
By setting the circumference at an appropriate position, the probability of occurrence of erroneous recognition can be sufficiently reduced. Further, in the color image recognition method according to the present invention, since the determination is made in consideration of not only the degree of coincidence of the pattern but also the characteristic value, if the characteristic value does not depend on the rotation of the input image, the characteristic value may be adopted. , Higher recognition rates can be achieved.

According to another aspect of the present invention, there is provided a recording medium storing a computer program for realizing the method according to the present invention when the recording medium is executed by a computer system. It is characterized by. Further, a recording medium according to claim 11 is characterized in that a computer program for realizing the method according to claim 9 is recorded by being executed by a computer system. That is, Claims 10 and 11
By executing a computer program recorded on each recording medium according to the present invention by a computer system, each color image recognition method according to the eighth and ninth aspects is realized.

[0016]

Embodiments of the present invention will be described below with reference to the drawings. A: First Embodiment FIG. 1 is a diagram showing a configuration of a color copying machine to which a color image recognition device according to a first embodiment of the present invention is applied. As shown in FIG. 1, a copying machine main body 10 is provided with an image input device 11 and an image output device 12.
The original on 1a is read and copied on the paper conveyed from the paper tray 10b. Operations such as start of copying and selection of a paper size are performed using the display 10d and the control panel 10e of the user interface 10c.

The image input device 11 includes the document scanning unit 1
1a, a wire 11b for driving the same, and a driving pulley 11c, and light from a document at the reading position is passed through a color filter in the document scanning unit 11a to red (R), green (G), and blue (B). And color image information corresponding to the lightness of each color is acquired using a CCD line sensor in the original scanning unit 11a, and the acquired color image information is converted into multi-gradation image signals R, G, and B. The image data is converted and output to the image processing apparatus 1 (described later). The CCD line sensor in the original scanning unit 11a is 400 dpi.
(Dot / inch) The original image is read at a resolution of 15.748 (dot / mm).

The image output device 12 is driven by a laser beam scanner 13 and a driving pulley 14a and is driven by a charger 14
and a photosensitive material belt 14 which is uniformly charged by c. The laser beam scanner 13 outputs image signals Y, M, C, and
K is converted into an optical signal, and the photosensitive material belt 14 is exposed to the optical signal via the polygon mirror 13b, the F-θ lens 13c, and the reflecting mirror 13d to form a latent image corresponding to the original image. Are developed by the developing units 14d corresponding to the respective colors of yellow (Y), magenta (M), cyan (C), and black (K), and a toner image of each color is formed on the photosensitive material belt 14. The developing device 14d has four colors (Y, M,
C, K), any one of the designated color toner images can be selectively developed.

On the other hand, the paper from the paper tray 10b is wound around the transfer belt 15 and rotates together with the transfer belt 15, and the color toner images on the photosensitive belt 14 are sequentially transferred to the paper. The transferred paper is discharged outside the color copier after the image is fixed by the fixing device 16. In addition,
The toner remaining on the photosensitive material belt 14 after the transfer of the toner image is removed by the cleaner 14b.

Here, the basic steps in the case of performing four-color full-color copying in the above-structured color copying machine will be described. FIG. 2 is a diagram schematically showing the timing of reading scanning in the image input device 11 and the timing of exposure in the image output device 12 in association with each other.
The right direction in the figure is the traveling direction of time. When performing four-color full-color copying, a total of four scans are sequentially performed on the original on the platen 10a by the original scanning unit 11a, and the image signals R and R are output in each of the four scans.
G and B are input. The image signals R, G, and B are supplied to the image processing apparatus 1, where they are subjected to well-known color correction processing, background removal processing, and the like, and are converted into color material information y, m, c, and k in the image output apparatus 12. Is done.

In the first reading scan, yellow color material information y is generated in the image processing apparatus 1 based on the image signals R, G, and B from the document scanning unit 11a, and laser light is generated based on the color material information y. The output unit 13a is driven, and the first exposure is performed. As a result, a latent image corresponding to yellow is formed on the photosensitive material belt 14. This latent image moves in the direction of the arrow in the figure with the rotation of the photosensitive material belt 14, and reaches the position of the developing device 14d. At the time of the first exposure, only the yellow portion in the developing device 14d is operable, and a yellow toner image is formed on the photosensitive material belt 14. Next, this yellow toner image is transferred onto a sheet of paper that is wound around the transfer belt 15 and rotates.

In the second reading scan, cyan color material information c is generated, a second exposure is performed based on the color material information c, and a latent image corresponding to cyan is formed on the photosensitive material belt 14. It is formed. This latent image moves in the direction of the arrow in the figure with the rotation of the photosensitive material belt 14, and reaches the position of the developing device 14d. At the time of the second exposure, only the cyan portion in the developing device 14d is operable, and a cyan toner image is formed on the photosensitive material belt 14. Next, the cyan toner image is transferred onto a sheet of paper wound around the transfer belt 15 and rotated. At this time, the yellow toner image has already been transferred onto the paper, and the transfer of the cyan toner image is performed in a state where the yellow toner image is aligned with the yellow toner image.

Thereafter, in the same manner as described above, the magenta and black toner images are sequentially superimposed and transferred on the paper in synchronization with the third and fourth reading scans, and a full-color image of four colors is formed on the paper. You. The paper on which the full-color image has been formed is peeled off from the photosensitive material belt 14, and is subjected to a fixing process of a toner image in the fixing device 16.

As described above, in the color copying machine having the above configuration, the laser beam scanner 13
Performs four reading scans. In some cases, a pre-scan for detecting the size of a document or the like may be performed prior to the four reading scans. Note that the four scanning operations are performed to form a full-color image without using a large-capacity memory. In order to form a full-color image by one reading scan, the image data obtained by one scan in the image input device 11 is stored in a page memory for each color, and the image output device 12 stores the image data of each color. Each time an image is sequentially formed, it is necessary to read image data from the page memory corresponding to each color. However, the amount of image data obtained by reading an original is enormous, and the amount of image data is extremely large in the case of color. Therefore, an extremely large-capacity memory is required as a memory for storing image data, resulting in high cost. There is a problem. On the other hand, when four reading scans are performed, image formation can be performed directly using image data obtained from the image input device 11 in synchronization with each scan, so that expensive page scanning is performed. There is an advantage that no memory is required.

In the color image recognition apparatus according to the present embodiment, the color copying machine having the above-described configuration detects a specific symbol indicating a copy-inhibited document from an input image (that is, image signals R, G, and B). Then, it is automatically determined whether or not the document is a copy prohibition document. In the present embodiment, circular symbols such as those illustrated in FIG. 9 are assumed as specific symbols. However, the present invention can be applied to detection of symbols having shapes other than circular by a simple design change. This is apparent from the following description.

FIG. 3 is a block diagram schematically showing the circuit system of the color copying machine shown in FIG. 1. In this figure, the same reference numerals are given to the same parts as in FIG. 1, and the description thereof will be omitted. As shown in FIG. 3, the image processing apparatus 1 receives image signals R, G, and B from the image input device 11, and outputs output image signals Y, M, and B obtained by converting the image signals R, G, and B.
C, K (color material information y, m, c, k) are output to the image output device 12
Output to In the image processing apparatus 1, the conversion from R, G, B to Y, M, C is performed by a well-known matrix operation represented by the following equation. Here, a11 to a33 are coefficients.

(Equation 1)

From the Y, M, and C obtained by the above equation, an inking signal k is obtained by the following equation, and k = α · min (Y, M, C). Under color removal is performed by the following equation. And final color material information y,
m and c are obtained. Here, α and β are coefficients.

(Equation 2) Outputs y, m, c, and k of the image processing apparatus 1 indicate yellow, magenta, cyan, and black after the undercolor removal.

As shown in FIG. 3, the image input device 1
The image signals R, G, B from 1 are supplied not only to the image processing apparatus 1 but also to the specific symbol detection circuit 2. The specific symbol detection circuit 2 outputs a symbol (hereinafter referred to as “specified”) indicating a copy prohibition in an input image represented by the image signals R, G, and B from the image input device 11 during prescan or main scan (four scans). Is determined, and if a specific symbol is included, a signal indicating that copying processing is prohibited is supplied to the image processing apparatus 1.

Here, a method of determining a specific symbol in the present embodiment will be described. In the present embodiment, regardless of the orientation of a document including a specific symbol, the orientation of the document hardly depends on the orientation of the document so that it is possible to accurately determine whether the specific symbol is included in the input image data. The characteristic value and the degree of coincidence are used. The characteristic values used in the present embodiment include the total number of ON pixels and the total number of ON / OFF inversions in a circle area preset according to the specific symbol, and the first circle preset according to the specific symbol. There are a total of four of the total number of ON pixels and the total number of ON / OFF inversions on the circumference. In addition, the degree of coincidence used in the present embodiment includes an image pattern (one-dimensional pattern) on a second circumference preset according to a specific symbol,
It is obtained by calculation based on a regular pattern set in advance according to a specific symbol. In the present embodiment, identification is performed based on a result of comparing each characteristic value with an allowable range of each characteristic value set according to a specific symbol, and a result of comparing the degree of coincidence with a preset allowable range. Since the symbols are detected, the specific symbols can be detected with high accuracy almost without depending on the orientation of the document.

The first circumference and the second circumference do not need to coincide with each other, but are preferably different from each other. Also, ON “total number of pixels” refers to the number of ON pixels in the area, and “ON / OFF inversion total number” refers to pixels that change from ON to OFF or OFF to ON in the area in the main / sub scanning direction. Refers to the number (number of times). Also, the "circular area"
This is an area surrounded by a circle of a specific diameter centered on the center position of the specific symbol determined from the size of the specific symbol to be detected, the reading resolution of the document, and the like. Note that the “specific diameter” is desirably set to the minimum length in which the specific symbol fits within the above-mentioned circular region, from the viewpoint of suppressing an increase in memory capacity and processing amount.

Further, "on the first circumference" is an area through which the circumference of the circle having the first radius passes, and "the first radius" satisfies the following relationship. (First radius) × 2 <(specific diameter) Further, the center point of the first circumference and the first radius are the first radius.
Is set so as to pass through a portion that clearly shows the characteristics of the specific symbol. The above relationship and setting policy are the same for the center point and the second radius of the second circumference.

Each of the above-described characteristic values represents the distribution characteristics of the pixels in the area as a numerical value. When the area is set on a specific symbol, as shown in Table 1, the scanning angle ( Even if the scanning direction of the original is different, the values are almost the same. Table 1 shows a measurement example of the total number of ON pixels and the total number of ON / OFF inversions in a circular area when a specific symbol having a circular shape is scanned at different angles. However, the reading resolution is 100 dpi, and 200 dots is about 12.5 m.
m.

[Table 1]

As is apparent from this table, the total number of ON pixels and the total number of ON / OFF inversions in the circular area do not change much even when the scanning angle changes. This is the same even if the specific diameter (circular diameter) changes. Therefore, even if the image to be scanned is rotating, the specific symbol can be detected by detecting the total number of ON pixels and / or the total number of ON / OFF inversions in the circular area. And 100
Since a specific symbol can be represented by a constant characteristic value even at a low resolution of dpi, color image data acquired at a high resolution of 400 dpi is converted into a low resolution of about 100 dpi by the image input device 11 and then detected. Even in this case, sufficient detection accuracy can be maintained. That is, the specific symbol circuit 2 can be realized with a small amount of hardware resources.

Although no measurement example is shown here, the total number of ON pixels and the total number of ON / OFF inversions on the first circle hardly change even if the scanning angle for a specific symbol changes. Therefore, even when noise at the time of reading is considered, it is estimated that the total number of ON pixels and the total number of ON / OFF inversions on the first circumference are substantially the same even when the scanning angle changes. Further, this is presumed to be the same even if the first radius is changed. In addition, the variance of the total number of ON pixels and the total number of ON / OFF inversions on the first circumference is expected to be extremely small compared to the variance of the total number of ON pixels and the total number of ON / OFF inversions in the circular area. Therefore, even if the image to be scanned is rotated, the total number of ON pixels on the first circumference and / or
Alternatively, a specific symbol can be detected by detecting the total number of on / off inversions. Further, utilizing the above-described property, the allowable range for the total number of ON pixels and the total number of ON / OFF inversions in the circular area is changed based on the fluctuations of the total number of ON pixels and the total number of ON / OFF inversions on the first circumference. In such a manner, the probability of occurrence of detection omission may be reduced even for an already output object having blurring and blurring.

For the detection of a specific symbol, the degree of coincidence is calculated from the image pattern located on the second circumference and the normal pattern, and the result of comparing the degree of coincidence with a predetermined allowable range is also obtained. Since it is used, it is possible to avoid erroneous recognition that an image other than a specific symbol is recognized as a specific symbol because the distribution characteristics of pixels in the region are similar.

Hereinafter, the configuration and operation of the specific symbol detection circuit 2 which detects each characteristic value, calculates the degree of coincidence, and detects a specific symbol using the characteristic values will be described with reference to FIG. Here, among the symbols exemplified in FIG. 9, it is assumed that a circular symbol in which a character “secret” is circled is a specific symbol to be detected. Also, here, the constituent color of the specific symbol (specific color)
Is a single color.

In FIG. 4, reference numeral 21 denotes a specific color detection circuit which receives image signals R, G, and B from the image input device 11 and outputs a predetermined color image from the image signals R, G, and B. Is extracted. That is, since the specific color detection circuit 21 extracts only the image of the constituent color of the specific symbol, the background color, the stain near the specific symbol, and the like are excluded from the image used for the determination. Therefore, the probability of occurrence of erroneous determination is reduced. Note that the configuration of the specific color detection circuit 21 is arbitrary. For example, threshold determination is performed on each of the input image signals R, G, and B, and “1” or “0” is output based on the determination results. You may make it. When a specific symbol has a plurality of constituent colors, the specific color detection circuits 21 are provided at least as many as the number of the constituent colors, and the image signals R, G, and B are input to the specific color detection circuits 21 in parallel. become. The specific color is set when the reference pattern is registered.

Reference numeral 22 denotes a first buffer memory for temporarily storing the output from the specific color detection circuit 21 as binary image data representing an image of a specific color, and is constituted by, for example, a semiconductor memory or the like. In the present embodiment, the image input device 11 has a resolution of 400 dpi in the main scanning direction and 40 dpi in the sub-scanning direction.
The original is scanned at a resolution of 0 dpi, and the image held in the first buffer memory 22 has the same resolution as the above resolution.

Reference numeral 23 denotes a resolution conversion circuit for converting the resolution of the binary image data output from the first buffer memory 21, and outputs the converted binary image data. That is, the resolution conversion circuit 23 includes the first buffer memory 2
From 1, the binary image data of the number of lines required in the subsequent circuit is read in parallel, and the read image data is
The process of converting the image data into image data of dpi × 100 dpi and outputting the image data is sequentially repeated. As described above, even if the resolution of the image data is reduced, the detection accuracy of the specific symbol can be maintained by using the characteristic values. Wear resources are being reduced. Although any conversion method can be adopted as the resolution conversion method, it is needless to say that a method capable of minimizing a decrease in detection accuracy due to the resolution conversion is desirable.

Reference numeral 24 denotes a second buffer for temporarily holding the binary image data after the resolution conversion output from the resolution converter 23.
, For example, a semiconductor memory or the like. The number of lines of the binary image data simultaneously held in the second buffer memory 24 is necessary to secure a predetermined circular area based on the size of the specific symbol to be detected, the reading resolution of the document, and the like. There are enough lines. This is not an indispensable condition. If it is not necessary to minimize the capacity of the second buffer memory 24, the number of lines of the binary image data simultaneously held in the second buffer memory 24 is required to secure the circular area. What is necessary is just to be more than the required number of lines.

Reference numerals 25, 26, 27, 28, and 40 denote a circle area ON pixel total number detection circuit, a circle area ON / OFF inversion number detection circuit, a first circumference ON pixel total number detection circuit, A circuit for detecting the total number of inversions on the circumference on / off and a second on-circle data coincidence calculation circuit for scanning the image data of a plurality of lines sequentially held in the second buffer memory 24 by the circle area I do. Note that the specific diameter of the circular area is set to a minimum length within which a specific symbol can fit within a circle of the specific diameter from the viewpoint of suppressing an increase in memory capacity and processing amount.

Next, the processing performed by each of the circuits 25 to 28 and 40 will be described below. The on-pixel count detection circuit 25 in the circular area detects the number of on-pixels in the circular area every time the circular area is moved, and the on / off inversion number detecting circuit 26 in the circular area moves the circular area. Each time it is performed, the number of pixels that change from on to off and from off to on in the circular area in both the main and sub scanning directions is detected. Each time the first circular on-pixel count detecting circuit 27 moves the circular area, the first circular on-pixel count detecting circuit 27 outputs the first
Detecting the number of ON pixels on a first circumference having a radius of
The first on / off inversion total number detection circuit 28 detects the total number of on / off inversions in the circumferential direction on the first circumference every time the circular area is moved.

The method of detecting each characteristic value by these circuits 25 to 28 is arbitrary. For example, the circuit 25 for detecting the total number of ON pixels in a circular area holds the detection result immediately before moving the circular area in the main scanning direction (or the sub-scanning direction), and stores the circular area in the main scanning direction (or the sub-scanning direction). The detection result may be obtained based on the pixel newly entering the circular area when moved to the pixel, the pixel coming out of the circular area and the immediately preceding detection result, or the circle Every time the area is moved, all pixels in the circular area may be examined to obtain a detection result.

Second on-circle data coincidence detecting circuit 27
Calculates the degree of coincidence between the image to be scanned and the specific symbol by performing an operation based on the image pattern on the second circumference having the second radius and the normal pattern with the center point of the specific symbol as the center. Is calculated. Here, details of the calculation of the coincidence will be described with reference to FIG. In FIG. 5, (a)
Is obtained by cutting out an image on the second circumference, and the normal pattern of (b-1) is an image pattern of a specific symbol, a center point of the specific symbol, and a second radius. Is set in advance based on. The normal pattern is obtained by superimposing a specific symbol on the second circumference and extracting an image pattern on the second circumference. A block with a coefficient of 1 is an ON pixel, and a block with a coefficient of -1. Indicates an off pixel, and a block having a coefficient of 0 indicates a pixel which cannot be said to be either. Such division is performed in consideration of blurring or blurring of a specific symbol, an environment at the time of inputting an image, and the like. Note that the coefficients that can be taken by each block of the normal pattern are not limited to 1, −1, and 0, and may be any value that accurately captures fluctuations due to blurring or blurring of a specific symbol, an environment at the time of image input, and the like. , 1,0.5,0, -0.
It may be 5, -1.

(B-2) to (b-17) are normal patterns obtained by shifting the normal pattern of (b-1) one block at a time, and correspond to the rotation of a specific symbol. Prepared in advance. The second on-circumference data coincidence detection circuit 27 first performs the following processing for each of the regular patterns. (1) For each pixel of the image pattern on the second circumference,
If the pixel is an ON pixel, the coefficient of the corresponding block on the normal pattern to be compared remains unchanged, and if the pixel is an OFF pixel, the sign of the coefficient of the corresponding block on the normal pattern is inverted. (2) The provisional coincidence is obtained by adding the coefficients of all the blocks.

When the provisional coincidences are obtained for all the normal patterns from (b-1) to (b-17), the second on-circle data coincidence detection circuit 27 determines the maximum of these provisional coincidences. The value is adopted as the degree of coincidence and output. As is evident from the above, the degree of coincidence also includes the degree of similarity with respect to the shape, unlike the feature amount. Therefore, a specific symbol can be detected with higher accuracy by combining with the detection using the feature amount.

It is also possible to realize a mode in which the values that can be taken by each block of the regular pattern are set to 1 and −1, and an intermediate value representing ambiguity is excluded. However, in order to obtain a sufficient effect in this mode, it is necessary to provide an environment in which it is not necessary to consider blurring or blurring of a specific symbol. For this reason,
This will be described with reference to FIG. In FIG. 6, (a) shows the second
Image pattern obtained by cutting out the image on the circumference of
(B) is a normal pattern set in advance based on the image pattern of the specific symbol, the center point of the specific symbol, and the second radius. The coefficient value of the white block is -1, and the coefficient value of the black block is 1. There is.

Here, when the detection / non-detection of a specific symbol is determined based on whether all pixels (all blocks) match, a result of "non-detection" is obtained for (a). . (A) can be regarded as an image pattern in which the line width of a part of the specific symbol is increased by one pixel and part of the line width is reduced by one pixel.
Since blurring and blurring in specific symbols and fluctuations due to the environment at the time of image input cannot be absorbed,
The result of "detection" cannot be obtained for the image pattern as shown in FIG.

In the case where the detection / non-detection is determined based on the degree of coincidence of the image pattern with the normal pattern,
The degree of coincidence with the image pattern of (a) matches the degree of coincidence with the image pattern of (c). The image pattern of (a) is considered to be partially thick or thin due to blurring or blurring of the line of the specific symbol, but the image pattern of (c) is not related to the line of the specific symbol. Have an ON pixel at a clear position. That is, the image pattern of (a) and the image pattern of (c) should be categorized, but they have the same degree of coincidence, so that they are treated the same. For this reason, the value that each block of the regular pattern can take is 1,
In order for the mode in which the intermediate value indicating ambiguity is excluded to be −1 to have a sufficient effect, it is necessary that the environment does not need to consider blurring or blurring of the specific symbol.

Referring again to FIG. 4, the determination circuit 29 determines whether or not the output results of the circuits 25 to 28 and 40 are within the respective allowable ranges set in advance according to the specific symbols. If the detection result is within the permissible range,
'1' is output in other cases and '0' is supplied to the image processing apparatus 1 in FIG. In the image processing device 1, when the value of the signal from the determination circuit 29 is “0”, normal copying processing is performed on the image signals R, G, and B from the image input device 11, and when the value is “1”. Output of the output signals Y, M, C, K corresponding to the image signals R, G, B to the image output device 12 is prohibited. As a result, copying of the copy-inhibited document is reliably prevented.

B: Second Embodiment Next, a color image recognition device according to a second embodiment of the present invention will be described. This embodiment is a modification of the color image recognition device according to the first embodiment so as to be able to handle a plurality of specific symbols. Hereinafter, only the portions different from the first embodiment will be described.

FIG. 7 is a diagram showing the configuration of a specific symbol detection circuit included in the color image recognition device according to the second embodiment of the present invention. In the configuration shown in FIG.
The same reference numerals are given to the same parts as those described above, and the description thereof will be omitted. The configuration shown in this figure is different from the configuration shown in FIG. 4 in that information on a specific symbol corresponding to the number and timing of reading scans is extracted from information on a plurality of specific symbols stored in advance, and based on the extracted information. The only difference is that a control circuit 50 for controlling the circuits 25 to 29 and 40 is provided. The control device includes, for example, a ROM (Read Only Memory) storing its own operation program,
A CPU (Central Processing Unit) that executes the operation program, and an EE that stores information on a plurality of specific symbols.
PROM (Electrically Erasable Programmable ROM)
And so on.

A plurality of specific symbol detection circuits 2 in the first embodiment may be provided in parallel, and a plurality of specific symbols may be detected by one scan. However, the required hardware resources are greatly increased. For this reason, in the present embodiment, a plurality of specific symbols are detected by using the operation characteristics of a general color copying machine without increasing the required hardware resources.
That is, in the four reading scans performed by the color copier, the specific symbols to be detected are sequentially changed, so that a maximum of four specific symbols are detected. Here, it is assumed that the four types of specific symbols to be detected are the four symbols shown in FIG. Further, since the second radius used in calculating the coincidence is desirably set so that the second circumference passes through a portion that clearly shows the feature of the specific symbol, as shown in FIG. , For each specific symbol. In FIG. 8, (a
The image pattern of -1) and the normal pattern of (b-1) correspond to the outermost circumference, and the image pattern of (a-4) and the normal pattern of (b-4) are the innermost. It corresponds to the circumference of the circumference.

In the present embodiment, the control circuit 50 controls each of the circuits 25 to 29, 40 based on information about a specific symbol corresponding to the number of read scans (1 to 4).
Done for Specifically, in each of the circuits 25 and 26, a center point corresponding to the specific symbol to be detected and a circular area based on the specific diameter are set, and in each of the circuits 27 and 28, a central point corresponding to the specific symbol is set. The first circumference is set based on the first radius, and the second data-on-circle calculation circuit 29 calculates the second circumference based on the center point corresponding to the specific symbol and the second radius. The circumference and the normal pattern are set, and the determination circuit 29 sets each allowable range or determination algorithm used for the determination. Therefore, a different specific symbol can be detected for each reading scan. Note that the sizes of the buffer memories 22 and 24 and the size of the storage area for storing image data, regular data, and the like are provided corresponding to the specific symbols requiring the largest storage area. .

The processing other than the above in each reading scan is the same as the processing according to the first embodiment, and the description thereof is omitted here. In the second embodiment, for the sake of simplicity, it is implicitly assumed that the constituent colors of the specific symbols and the conversion magnification of the resolution are common, but the control circuit 50 uses , 23 can be handled even if the constituent colors of each specific symbol and the conversion magnification of the resolution are different.

C: Supplement In each of the above-described embodiments, the image data input to the specific symbol detection circuit 2 has been described as the image signals R, G, and B. However, the present invention is not limited to this. By changing the processing content (and configuration) of the detection circuit, it is possible to cope with image data in different color spaces such as CMYK and L * a * b *. In each of the above embodiments, the resolution is converted after the specific color image is extracted by the specific color detection circuit 21. However, the resolution conversion circuit 24 is provided before the specific color detection circuit 21 to reduce the resolution. A specific color may be detected to improve the detection accuracy.

Although not explicitly shown in each embodiment, the first radius and the second radius may be input by a human operating the instruction means, or may be automatically input based on a specific symbol. You may ask for it. Needless to say, each characteristic value on the first circumference has an allowable range,
The normal pattern for the image pattern on the circumference of is obtained manually or automatically after the first radius and the second radius are set, respectively.

Further, in each embodiment, when the value of the signal of the determination circuit 29 is “1”, the output of the image signal to the image output device 12 is prohibited, so that the copy of the copy prohibited document is prohibited. However, if the purpose is achieved only by preventing normal copying, such as to prevent counterfeiting of securities, only specific image signals (for example, any of image signals C, M, Y, and K) should not be output. Alternatively, a specific image signal may be a signal for constantly exposing, so that the copy image output on the transfer paper may be filled with a specific color.

Further, in each of the above-described embodiments, the configuration color of one specific symbol is described as one color, but the present invention is not limited to this. For example, when the constituent colors of a specific symbol are “red” and “yellow” and the constituent colors of other specific symbols are “red” and “blue”, the characteristic value detection circuit and the determination circuit are regarded as the specific symbol. This can be dealt with by providing for each combination with a color and making a comprehensive determination in a newly provided comprehensive determination circuit. In the above case,
The circuit for extracting the “red” pixel may be configured to be shared in the detection processing of both specific symbols.

In order to reduce the hardware resources, the first on-circle ON pixel total number detecting circuit 27 and the first
Of the on / off inversion total number detection circuit 28 may be provided, or any of the on-pixel total number detection circuit 25 in the circular area and the on / off inversion number detection circuit 26 in the circular area may be provided. Alternatively, only one of them may be provided. Further, only the second on-circle data coincidence calculation circuit 40 may be provided, or the second on-circle data coincidence calculation circuit 40 obtains the coincidence for a plurality of circumferences having different radii. You may do so. Such a modification can be appropriately performed according to the requirements for hardware resources and determination accuracy.

In each of the embodiments, it is assumed that the present invention is applied to a color copier. However, the present invention is not limited to this, and color image data of a bit map such as a facsimile and an image editing device can be used. The present invention is applicable to any color image recognition device that inputs and outputs an image represented by input color image data or image data obtained by processing the image data. The method of inputting an image is arbitrary, and color image data supplied from a computer network or the like may be input.

Further, the color copying machines according to the embodiments can be applied to purposes other than those described above. For example, a process of re-inputting a copy to obtain an output (re-copying: co
It can be modified to a color copier for the purpose of improving the quality of an output image in py to copy. Specifically, a specific symbol indicating a copy mode (“photo mode”, “map mode”, etc.) is detected from a re-input copy, and parameters related to copying are set in accordance with the detected specific symbol. For example, humans can be excluded from the copying mode selection process, and the quality of re-copying can be improved with a small amount of hardware resources. That is, by appropriately changing the processing performed by the determination circuit 29, it can be applied to various uses.

Note that the technical scope of the invention according to the present application is not limited to the above embodiment, and various changes are made to the above embodiment to carry out the invention described in the claims. be able to. It is obvious from the description of the claims that such an invention belongs to the technical scope of the invention according to the present application. For example, a recording medium on which a computer program for realizing the operation of the specific symbol detection circuit 2 is recorded belongs to the technical scope of the present invention.

[0064]

As described above, according to the present invention,
A specific symbol can be detected from the input color image with less hardware resources and higher accuracy than when a general pattern matching method is applied.
Further, even when the direction of the specific symbol does not match the direction of the image provided with the specific symbol, the specific symbol can be recognized with high accuracy.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a color copying machine to which a color image recognition device according to a first embodiment of the present invention is applied.

FIG. 2 is a diagram schematically showing timings of reading scanning in an image input device 11 and exposure in an image output device 12 of the copying machine.

FIG. 3 is a block diagram schematically showing a circuit system of the copying machine.

FIG. 4 is a block diagram showing a configuration of a specific symbol detection circuit 2 in the copying machine.

FIG. 5 is a diagram illustrating a process of detecting a degree of coincidence in the specific symbol detection circuit 2;

FIG. 6 is a diagram illustrating a detection process different from the detection process.

FIG. 7 is a block diagram showing a configuration of a specific symbol detection circuit in a color copying machine to which a color image recognition device according to a second embodiment of the present invention is applied.

FIG. 8 is a diagram illustrating a process of detecting a degree of coincidence in the specific symbol detection circuit.

FIG. 9 is a diagram illustrating an example of a specific symbol.

[Explanation of symbols]

21: specific color detection circuit, 22: first buffer memory, 2
3: resolution conversion circuit, 24: second buffer memory, 25
... A circuit for detecting the total number of ON pixels in a circular area, 26 a circuit for detecting the number of ON / OFF inversions in a circular area, 27 a circuit for detecting a total number of ON pixels in a first circle, 28. Detection circuit, 29: determination circuit, 40: second circumference data coincidence degree calculation circuit

Claims (11)

[Claims] 1. A color image recognition apparatus for receiving a color image signal and recognizing a specific symbol from an image represented by the color image signal, wherein a specific color extracting unit extracts an image of a component color of the specific symbol from the image. Characteristic value detecting means for detecting a characteristic value relating to the specific symbol from an image in the area while scanning the image extracted by the specific color extracting means in an area having a size corresponding to the specific symbol; Among the images extracted by the extraction means, the image in the area and the specific symbol are determined based on an image pattern representing an image located on a circumference centered on the center pixel of the area and a predetermined regular pattern. A coincidence calculating means for calculating a coincidence with: a characteristic value detected by the characteristic value detecting means; a coincidence calculated by the coincidence calculating means; Color image recognition, comprising: determination means for determining the presence or absence of the specific symbol based on a gender value and an allowable range preset for the degree of coincidence, and outputting a determination result as a recognition result. apparatus. 2. The method according to claim 1, wherein the degree of coincidence calculating means determines an operation to be performed on a value of a block corresponding to each pixel of the normal pattern, based on a state of each pixel of the image located on the circumference. 2. The color image recognition device according to claim 1, wherein the degree of coincidence is calculated based on a result of the calculation of the calculation content. 3. The apparatus according to claim 1, further comprising an instruction unit for indicating the diameter of the circumference, wherein the regular pattern is a pattern corresponding to the diameter indicated by the instruction unit. Color image recognition device. 4. The color image recognition device according to claim 1, wherein the area is a circular area. 5. The color image recognition device according to claim 1, wherein the area is an annular area. 6. The color image recognition apparatus according to claim 1, further comprising control means for selecting one symbol from the plurality of symbols and using the selected symbol as a specific symbol to be recognized. 7. A color image recognition device that receives a color image signal and recognizes a specific symbol from an image represented by the color image signal, wherein a specific color extracting unit extracts an image of a component color of the specific symbol from the image. And, among the images extracted by the specific color extracting means, an image pattern representing an image located on a circumference centered on a center pixel of an area having a size corresponding to the specific symbol, and a predetermined normal pattern. A coincidence calculating means for calculating a degree of coincidence between the image in the area and the specific symbol based on the degree of coincidence calculated by the degree of coincidence calculation and an allowable range preset for the degree of coincidence. A determination unit configured to determine the presence or absence of the specific symbol based on the determination result and to output a determination result as a recognition result. 8. A color image recognition method for receiving a color image signal and recognizing a specific symbol from an image represented by the color image signal, wherein a specific color extracting step of extracting an image of a component color of the specific symbol from the image A characteristic value detecting step of detecting a characteristic value related to the specific symbol from an image in the area while scanning the image extracted in the specific color extracting step in an area having a size corresponding to the specific symbol; Of the images extracted in the color extraction step,
Matching degree calculation for calculating the matching degree between an image in the area and the specific symbol based on an image pattern representing an image located on a circumference centered on the center pixel of the area and a predetermined regular pattern And the characteristic value detected in the characteristic value detecting step, the degree of coincidence calculated by the degree of coincidence calculation means, and the allowable range preset for the characteristic value and the degree of coincidence. A determination step of determining the presence or absence of a specific symbol and outputting the determination result as a recognition result. 9. A color image recognition method for receiving a color image signal and recognizing a specific symbol from an image represented by the color image signal, wherein a specific color extracting step of extracting an image of a component color of the specific symbol from the image And an image pattern representing an image located on a circumference centered on a center pixel of an area of a size corresponding to the specific symbol among the images extracted in the specific color extracting step, and a preset normal pattern. A matching degree calculating step of calculating a matching degree between an image in the area and the specific symbol based on the matching degree; a matching degree calculated in the matching degree calculating step; and an allowable range preset for the matching degree A determination step of determining the presence or absence of the specific symbol on the basis of the above, and outputting the determination result as a recognition result. 10. A recording medium having recorded thereon a computer program for executing the method according to claim 8, when the computer program is executed by a computer system. 11. A recording medium on which is recorded a computer program for executing the method according to claim 9 when executed by a computer system.