JPH07226839A - Picture processor - Google Patents

Abstract

PURPOSE:To effectively prevent the unauthorized use of a reproduced picture by recognizing a mark by providing a means for recognizing the mark in a picture being a processing object and varying a picture processing as against the picture being the processing object in accordance with the presence or absence of the recognition of the mark. CONSTITUTION:A feature extraction means 2 extracts picture feature information from input picture data which is read by a picture read means 1 in a picture element unit, and it is written into a feature buffer 3. A mark registration means 4 previously registers picture feature information of the mark M having the size straddling over plural picture elements and consisting of prescribed picture feature information. A mark judgement means 5 retrieves information in the buffer 3 by a retrieval window 6 which is divided into the same size as the mark M or size smaller than it. When the information pattern of the mark M registered in the means 4 is checked and it is matched with an information pattern in the window 6 or they are approximated, the mark M is judged to exist. When the mark M is judged, a mark picture processing means 8 directly conducts a mark picture processing which the mark M means.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for performing image processing on an image captured by a scanner or the like for each pixel unit and outputting the image, and in particular, including a predetermined mark in an image to be processed. The present invention relates to an image processing device capable of realizing appropriate image processing by recognizing a mark.

[0002]

2. Description of the Related Art Recently, with the rapid improvement in the performance of color copying machines, it has come to be possible to obtain a copy quality that is very close to that of the original. Under these circumstances, crimes involving misuse of color copiers are occurring.

Further, confidential documents in a company can be easily copied by using a copying machine, so that there is still concern about leakage of confidential information.

[0004]

In order to solve such an unexpected situation, the image processing which makes it impossible to copy the object which is not desired to be copied or which can be subjected to a special image processing. The advent of devices is desired. The present invention has been made to solve such a technical problem, and an image processing apparatus capable of accurately recognizing a mark in an image to be processed and reliably preventing the reproduction image from being misused. Is provided.

[0005]

That is, the present invention is
As shown in FIG. 1, a feature extracting unit 2 for extracting image feature information from the input image data read by the image reading unit 1 on a pixel-by-pixel basis, and the image feature information extracted by the feature extracting unit 2 for each pixel. A feature buffer 3 to be written for each unit,
A mark registration means 4 for pre-registering image feature information of a mark M having a size across a plurality of pixels and consisting of predetermined image feature information, and a search window 6 divided into the same size as or smaller than the mark M. The image feature information in the feature buffer 3 is sequentially searched, and whether the image feature information pattern in the search window 6 matches or is similar to all or part of the image feature information pattern of the mark M registered in the mark registration means 4. Mark determination means 5 for determining whether or not the marks M are all or a part of the marks M if they match or are similar to each other is provided.

In such an image processing apparatus, when it is necessary to use the position information of the mark M in the image to be processed, the mark determination means 5 determines that it is the mark M, as shown in FIG. It suffices to provide the mark position storage means 7 for storing the position of the created image area. If such mark position storage means 7 is provided, the positional relationship between the two or more marks M can be determined by using the mark position information, or the mark image processing means 8 can use the mark position storage means 7. It is possible to execute a specific mark image process (for example, black painting) on the mark image area based on the mark position information.

When the mark determination means 5 determines that the mark is M, the mark image processing means 8 executes a specific mark image processing (for example, copy prohibition processing) that the mark M means. You may

In such a technical means, the feature pixel extracting algorithm by the feature extracting means 2 may be appropriately selected, but from the viewpoint of increasing the extraction accuracy of the feature pixel, the position around the pixel of interest is determined. It is preferable to adopt a method of referring to a plurality of pixels that are selected and extracting the characteristic pixel of the target pixel. Further, as the feature buffer 3, a minimum required capacity may be secured according to the feature pixel extraction method by the feature extraction means 2. The expression format of the characteristics of the characteristic pixel written in the characteristic buffer 3 may be appropriately selected as long as it is predetermined. Further, from the viewpoint of further enhancing the characteristic pixel extraction accuracy, it is determined whether or not the characteristic pixel written in the characteristic buffer 3 is noise, and if it is noise, the property of the characteristic pixel is deleted from the characteristic buffer 3. It is preferable to add noise elimination means. In this case,
The algorithm for determining whether or not the characteristic pixel is noise may be appropriately selected.

Further, the mark M in the image to be processed, which is judged by the mark judging means 5, may be appropriately selected as long as it can be clearly distinguished from a normal image. In this case, it is preferable that the mark M is printed such that the image feature information pattern is generally the same as or similar to a small number of combinations or ratios thereof. Then, even if the mark M is rotated with respect to the image reading means 1 by an arbitrary angle, at least one color that does not change the number or the ratio of the matching image feature information patterns in the search window 6 is included. It is possible to determine the mark M without being affected by the rotation angle of. On the other hand, when the mark M is rotated with respect to the image reading means 1 by an arbitrary angle, if at least one color in which the number or the ratio of the matching image characteristic information patterns changes is included in the search window 6, the change occurs. It is possible to specify the rotation angle of the mark M based on the degree. Also,
The mark M is not necessarily limited to including only one piece of information, and may include a plurality of pieces of information. For example, the mark M may be divided into a peripheral portion and a central portion, and different information may be recorded in the inner area and the outer area. Furthermore, in order to effectively avoid the situation where the mark M is erased when the mark M is added to the image to be processed, the mark M is mixed with the surrounding pattern or color so that the mark M can be recognized by the human eye. It is preferable to make it difficult to notice.

Further, in the type in which the search window 6 is divided into smaller than the size of the mark M, in order to quickly perform the determination process of the mark M, the mark determination means 5 searches within the image area corresponding to the mark M. Window 6
Are sequentially moved to each divided area of the search window 6,
The image feature information pattern in the search window 6 is the mark M.
It is preferable to end the determination process in the image area corresponding to the mark M when it is determined that it is not a part of the above, and to shift to the next process.

Further, when the inclination angle of the original 9 with the mark M is also determined, for example, an arc-shaped orientation determining component different from other mark portions existing only in at least a partitioned area of the mark M is determined. Ma may be provided so that the mark determination means 5 recognizes at which position the orientation determination component Ma is present.

Furthermore, although the ink printed on the mark M usually fades, from the viewpoint of accurately judging the mark M even if the color fades, the feature extracting means 2 can be used as a mark of the mark M. The characteristic information used by the mark determination means 5 is obtained by using a direction component having a low sensitivity to the fading change direction (a direction component intersecting at right angles with the direction in which the ink printed on the mark M is fading in the color space). It should be extracted.

[0013]

According to the technical means as described above, the feature extraction means 2 extracts the image feature information from the input image data read by the image reading means 1 on a pixel-by-pixel basis. It is written in the feature buffer 3 every time.
On the other hand, the mark registration means 4 pre-registers the image feature information of the mark M having a size across a plurality of pixels and including predetermined image feature information. Then, the mark determination means 5 sequentially searches the image feature information in the feature buffer 3 in the search window 6 divided into the same size as or smaller than the mark M, and the image feature information pattern in the search window 6 is marked. It is checked whether or not all or part of the image feature information pattern of the mark M registered in the registration means 4 is the same or similar, and if they are the same or similar, the mark M is determined.
Is determined to be all or a part of. When the mark M is determined in this way, the mark image processing means 8 directly executes the mark image processing that the mark M means, or
Alternatively, specific mark image processing is executed on the mark image area based on the mark position information of the mark position storage means 7.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below based on the embodiments shown in the accompanying drawings. Example 1 FIG. 2 is an example of a color image forming apparatus to which the present invention is applied. In the figure, reference numeral 10 is a scanner for reading a color original document, and after the input image data read by the scanner 10 is converted into polar coordinate display of a uniform color space that has been subjected to appropriate preprocessing such as background removal, Image processing unit 2
Sent to 0.

In the image processing unit 20 according to this embodiment, reference numeral 21 is an image buffer for temporarily holding the image data (quantized continuous tone data) from the scanner 10, for example, N (positive integer) rows. In the case of performing the processing one by one, it is necessary to have a storage capacity capable of holding at least N + 2 rows of multivalued image data.

Reference numeral 22 is a feature extraction unit for extracting the feature information of the image based on the image data, and 23 is a feature buffer for writing the feature information of the image extracted by the feature extraction unit 22. In addition, the individual image processing unit 24 (specifically, 2
4 (1) to 24 (n)) determines whether or not the processing by itself is necessary based on the information in the feature buffer 23, and if necessary, how strong the processing is. The image of the image buffer 21 is processed according to the decision. This process is repeated in order from the individual image processing unit 24 (1) to the individual image processing unit 24 (n), and finally the image buffer 21 contains an image that has been appropriately subjected to the necessary processing.
In this embodiment, the individual image processing unit 24 (1) is a mark recognition processing unit, and the other individual image processing units 24 (2) ... Are edge enhancement processing, color compression, noise removal, and the like.

Thereafter, the processing is transferred to the image output unit 30,
Commonly known color correction 31 and density correction 3
After each processing such as 2, binarization 33, etc., an image is output.

Next, the configuration of the image processing unit 20 will be described in more detail with reference to FIG. 3, centering on the mark recognition processing unit 24 (1). In the figure, H from the image buffer 21
Three image signal components of VC (Munsell color system) (brightness,
The chromaticity, the saturation) are transmitted, and the feature extraction unit 22 includes a feature pixel extraction unit 221 that extracts a feature pixel for a signal obtained by linearly combining three image signal components.

FIG. 4 shows a specific example of the characteristic pixel extraction unit 221 according to this embodiment. In the figure, the characteristic pixel extraction unit 2
Reference numeral 21 denotes an image signal synthesizing unit 41, an edge characteristic pixel extracting unit 42 for extracting edge characteristic pixels corresponding to an edge pixel portion having a large image change, and a gradient for extracting a gradient characteristic pixel including a pixel portion having a gradient in the image signal. Feature pixel extraction unit 43
It has and. In this embodiment, the image signal synthesizing unit 41 appropriately switches the coefficient when the image signal components are linearly synthesized for each type of the feature pixel to be selected, and when extracting the edge feature pixel, the brightness, The coefficients of chromaticity and saturation are α1, β1, and γ1, and when the density gradient feature pixel is extracted, the coefficients of lightness, chromaticity, and saturation are α2,
Let β2 and γ2. In this embodiment, when importance is attached to the degree of influence of lightness, α1, α2 = 0.6, β1, β2 = 0, γ
1 and γ2 = 0.4, and when importance is attached to the change in hue, α1, α2 = 0, β1, β2 = 0.6, γ1, γ2 = 0.
Use properly as 4.

Then, the characteristics of the characteristic pixels extracted by the characteristic pixel extracting section 221 are written in the characteristic buffer 23. As shown in FIG. 3, the characteristic buffer 23 has the edge characteristic pixel extracting section 42 and the gradient characteristic pixel. Corresponding to the pixel extracting unit 43, N + 2 row line buffers 231 and 232 each having a storage capacity of N + 2 rows are provided. Reference numeral 222 denotes a noise deleting unit that determines whether or not the characteristic pixel written in the characteristic buffer 23 is noise and, if it is noise, deletes the noise.

Next, the mark recognition processing unit 24 (1) has a mark discriminating unit 241. The mark discriminating unit 241 registers the characteristic information of the edge and the gradient written in the line buffers 231 and 232 as the registered mark information. The image to be processed is classified by comparing with the characteristic information of the registration mark stored in the storage unit 242 in advance. Here, when the mark discrimination unit 241 discriminates that a part of the image is identical or similar, the registration mark candidate image address buffer 243 has an address (position on the image) on the image where this mark is registered. Is to be written.

Next, the image feature extraction process, the mark recognition process, and other individual image processing steps adopted in the color image forming apparatus according to this embodiment will be described with reference to the flow chart of FIG. In the figure, first, the feature buffer 2
The pointer 3 is moved to a predetermined position (step 1), and the characteristic pixel extracting unit 221 selects the characteristic pixel while scanning the image of the required number N of rows input from the image signal synthesizing unit 41. If so, the type symbol of the characteristic pixel is recorded in the corresponding position of the characteristic buffer 23 (step 2). Next, it is determined whether or not the characteristic pixel selected by the noise removing unit 222 is noise depending on whether or not there is another adjacent characteristic pixel. If not, the type symbol of the characteristic pixel is deleted from the corresponding position of the characteristic buffer 23 (step 3). After that, after the processing of the input image is moved to the Nth row ahead (step 4), the noise-removed characteristic pixel is used to selectively perform the mark recognition processing on the processing target pixel. And
The processed pixels are written in the image buffer 21. If there is still another image processing, the processing is continued (step 5). Then, it is determined whether or not all the rows of the input image have been scanned (step 6), and the processing from step 1 to step 5 is repeated until the processing of all the rows is completed.

Further, this processing will be described more specifically.
Now, assuming that the input image is stored at 400 spi and the number of gradations is 256, that is, one pixel is stored at 8 bits, the case of processing the image every two rows (N = 2) will be described. As shown in FIG. The capacity of the buffer 21 and the feature buffer 23 requires at least N + 2 rows, that is, 4 rows. At this time, the extracted feature information is already written in the top row of the feature buffer 23, and the noise removal is also completed. Further, the extracted feature information is written in the second line, but the noise removal is not completed. Further, the third line and the fourth line contain the used characteristic information that has already been extracted and used by the mark recognition processing unit 24 (1) or copied to another storage area. Therefore, the newly extracted feature information will be written in the third and fourth lines.

Next, a method of selecting characteristic pixels by the characteristic pixel extraction unit 221 will be described. Now, as shown in FIG. 7, it is assumed that the pixel of interest a is at the center of the 3 × 3 pixel matrix with respect to the pixel of 3 × 3 matrix, and the peripheral pixels are b, c, d, e, f,
As g, h, and i, the calculation according to the following equations 1 (1) and (2) is performed.

[0025]

[Equation 1]

In Expression 1 (2), Sum represents the difference between the pixel a and the average of the pixels b to i around the pixel a.
That is, if the value of Sum is small, it is not different from the surrounding pixels and cannot be a characteristic pixel. A specific example is shown in FIG.
It shows in (b). Here, only the brightness information is extracted from the image signal for description. FIG. 8A shows a substantially uniform lightness distribution. In this case, Sum = −4.375, which is small as expected. FIG. 8B is an example having a lightness gradient that rises to the right. In this case, Sum = 1.75, which is also a small value. Although it is detected as an edge by a general edge filter, since it is necessary to perform more detailed processing here, the rising edge and the falling edge (beginning of the falling edge) are distinguished, and a gentle gradient is determined as an edge feature pixel. Will not judge. However, when it is desired to use the gradient information in the subsequent individual image processing, the following mathematical expression 2 may be performed.

[0027]

[Equation 2]

Even if Sum is small and it is not determined to be a characteristic pixel, if Abs is large in Formula 2, there is a gradient.

The following expression 3 (1) is adopted as the judgment condition of the edge characteristic pixel. Also, even if Sum> CC1
Even if it satisfies, it may be noise. The noise determination is also performed by the noise removal unit 225 in the next stage, but the determination is also performed here by a simple method. When Sum is extremely large, only the pixel of interest has low or high brightness, and is likely to be noise. Therefore, if the condition of Expression 3 (2) is satisfied, it is determined to be noise. Therefore, Sum
If is CC1 <Sum <CC2, it is determined that the pixel a is an edge feature pixel. Further, if the gradient feature pixel is also desired to be detected, it is determined that there is a gradient when the pixel of interest satisfies the condition of the following expression 3 (3).

[0030]

[Equation 3]

Next, it is determined in which direction the edge feature pixel and the gradient feature pixel have the feature, that is, in which direction the brightness change of the pixel a is large. Therefore, Ma
x = max {| B |, | C |, | D |, | E |, | F
|, | G |, | H |, | I |} is determined to determine the direction of brightness change, and the sign of the difference between the average value of the pixels in that direction and the pixel of interest a is used to determine the positive or negative of the slope. For example, FIG. 8 (c)
In the case of, the direction is e +. This is expressed here as ↓ + for convenience, and is recorded in the feature buffer 23. In the case of gradient feature pixels, the directions are → +, ←-
Means the same thing, so only the + direction is adopted,
An underline is added to indicate that this represents a slope and is displayed as → .

FIG. 9 shows symbols corresponding to changes in the brightness of the image. In the figure, two characteristic pixels → +, ← −
It can be seen that there is a large change between the two pixels where is facing each other. Also, if the gradient is gentle, it can be seen that some pixels have gradient symbols in succession, and in the case of a relatively sudden change, at both ends or
Characteristic pixels → + and ←-should appear in either one. Actually, CC1, CC1 ', and CC1', so that the image and the marks of the edge feature pixel / gradient feature pixel correspond well in this way.
CC2 needs to be determined.

Next, the details of the noise removing section 222 will be described. Now, in FIG. 6, since the noise determination is finished only for the first line of the feature buffer 23, the data of the first line is used when processing the second and third lines. Here, there is no problem in the case of performing the noise determination after the feature extraction for two rows is completed, but in the case of performing the feature extraction and the noise determination for each pixel, the pixels are processed in the order shown in FIG. There is a need to continue to. That is, since it is necessary to finish the feature extraction of the surrounding pixels for noise determination, it is necessary to finish the feature extraction up to pixel D5 in order to perform noise determination for pixel * in FIG. This means that the feature extraction must be completed up to pixel D6.

If the pixel of interest is extracted as an edge feature pixel, it is checked whether or not there is a feature pixel having a directional component similar to the surrounding pixels, as shown in FIG. If not, it is judged as noise and the edge feature pixel mark is deleted. Here, the similar direction component is a direction in which the edge feature pixels have the same sign and are within ± 80 ° from the direction of the target pixel. For example, if the pixel of interest is ↓-, it is ↓-, ↓ (tilt 45 ° to the right)-, and ↓ (tilt to the left 45 °)-. Therefore, the pixel of interest in FIG. 11 is an edge feature pixel rather than noise. Further, as shown in FIG. 12, even if the extracted edge feature pixel is isolated, there is a pixel showing a gradient in the direction in which the edge feature pixel is facing, and if the edge feature pixel is +, it is similar. The gradient of the direction,
In the case of −, a method of determining that the pixel of interest is truly a feature pixel even when the gradient in the opposite direction is expressed is adopted.

As described above, noise pixels are removed,
After organizing the information of the characteristic pixels, the mark recognition processing unit 24
Starting from (1), each individual image processing unit 24 performs desired image processing on the input image using the information of the characteristic pixel. Depending on the processing content of the individual image processing unit 24, it is necessary to configure the characteristic pixel extraction unit 221 that satisfies each request.

Next, the mark recognition processing will be described in detail. In this embodiment, as shown in FIG. 13, 80 × 80
A mark M having a pixel size (in this embodiment, a thick line-shaped quarter arc inside, a thin line circle outside the circle, and a thick broken circle outside the circle) is a recognition target, and FIG. , The mark M is divided into four small areas (one small area is 40
(× 40 pixels) R1 to R4 are divided and recognized.
Therefore, in order to perform the mark recognition processing for one line of the image area, the feature buffer 23 needs to be buffered for 40 lines, and for the noise determination, 42 lines must be buffered. After that, the feature extraction process and the mark recognition process are alternately repeated for each line.

As shown in FIG. 15, the feature extraction processing is performed while moving the 40 × 40 search window 50 on the feature buffer 23 from left to right. The flow of this feature extraction processing is shown in FIG. First, the case where the address in the upper left corner of the window is compared with the address stored in the registration mark candidate image address buffer 243 and the value is different (step 1) will be described. At this time, the window 50 will search the upper left area of the mark M as shown in FIG. First, the characteristics displayed in the window are summed up for each type, and the value is calculated as a ratio to the totalized value (hereinafter, this value will be referred to as a characteristic ratio). More specifically, this process starts up,
Assume that the sum of the falling and slope features are as shown in Table 1 below.

[0038]

[Table 1]

In this case, I1 in the following equation (1) is a total value obtained by summing the rising features by type, I2 is a total value obtained by summing the falling features by type, and I3 is a gradient feature by type. Equation (4) means the feature ratio for each type of feature.

[0040]

[Equation 4]

Then, each feature ratio is compared with the feature ratio calculated for the mark M in advance, and if the value does not satisfy the condition 1 of the following expression 5, the window is shifted to the right by one and the same. The process is repeated (steps 2, 3,
7). If the value satisfies the condition 1, the window is moved by 40 pixels and the same process is performed, that is, the features displayed in the window are summed for each type and the feature ratios are compared. Therefore, when the total value for each type satisfies the following condition 2, the window for 39 pixels is returned, and the position (address) of the pixel adjacent to the lower left at that position is stored in the registration mark candidate image address buffer 243 and processed. Is repeated from the beginning (steps 4, 5, and 7). If the condition 2 is not satisfied, the window for 39 pixels is returned and the process is returned to the initial process (steps 4, 6 and 7).

Next, the case where the address in the upper left corner of the window is compared with the address stored in the registration mark candidate image adlay buffer 243 and the values match (step 1) will be described. At this time, it is known that the upper area (see 51 in FIG. 17) of the overlapping images of the windows 50 matches the upper half of the registration mark.
The window 50 searches the lower left area of the mark M as shown in FIG. As before, the feature ratio of the image in this window is compared with the pre-calculated feature ratio of the lower left region of the registration mark, and the condition 3 of Equation 5 which is the determination condition of the lower left region of the registration mark is obtained. Compare (step 8). If the condition 3 is not met, it is determined that this image is not a registration mark, the window returns to the first processing at this position, and the process is compared with the condition 1 (step 2). If they match, the window is shifted to the right by 40 pixels, and the feature ratio of the image in the window is compared with the pre-calculated feature ratio of the lower right region of the registration mark, and the lower right region of the registration mark is compared. It is compared with the condition 4 of the equation 5 which is the judgment condition of (step 9). If they match, that is, if Condition 4 is satisfied, it is determined that the image is a registration mark (step 10). If they do not match, the pixel is returned by 40 pixels, the process returns to the beginning, and comparison with Condition 1 is started (steps 11 and 2). In this way, the mark recognition process is continued.

If the process returns to some other process at the time of recognizing the mark, the process is interrupted at this time, but if it is judged how many registration marks are in the image, This process is continued until the end of the target image area.

[0044]

[Equation 5]

In Equation 5, the upper and lower limit values of each feature ratio indicate the similar range of each feature ratio of each region (upper left region, upper right region, lower left region, lower right region) with respect to the registration mark. In addition, in this embodiment, in conditions 1 to 4,
The condition is that all the inequalities are satisfied, but any one of the typical ones or a plurality of predetermined combinations may be satisfied.

Example 2 In Example 1, the method of extracting the edge feature and the gradient feature by using the 3 × 3 mask is shown. However, when the image has a lot of noise, it is better to use the larger mask. In order to handle line drawings and the like, it is appropriate to adopt a process that can also extract straight lines. This embodiment is provided with an image processing unit 20 that meets such a request.
The details of the above are shown in FIG. Unlike the first embodiment, the image processing unit 20 includes a feature pixel extraction unit 221 that extracts a feature related to a line in addition to an edge and a gradient. Therefore, the feature buffer 23 also relates to an edge, a slope, and a line. Three N + 2 line buffers 23 into which features are written
1 to 233 are provided.

FIG. 19 shows the detailed arrangement of the characteristic pixel extraction unit 221 according to this embodiment. In the figure, the feature pixel extraction unit 221 includes an image signal synthesis unit 41, an edge feature pixel extraction unit 42 that extracts an edge feature pixel corresponding to an edge pixel portion having a large image change, and a gradient including a gradient pixel portion. A gradient feature pixel extraction unit 43 for extracting feature pixels,
A line characteristic pixel extraction unit 44 including a linear pixel portion is provided. In this embodiment, the image signal synthesizing unit 41 appropriately switches the coefficient when the image signal components are linearly synthesized for each type of the feature pixel to be selected, and when extracting the edge feature pixel, the brightness, Chromaticity,
Saturation coefficients are α1, β1, γ1, and when extracting density gradient feature pixels, brightness, chromaticity, and saturation coefficients are α2, β2,
γ2, and when extracting line characteristic pixels, the coefficients of lightness, chromaticity, and saturation are α3, β3, and γ3. In this embodiment, when importance is attached to the degree of influence of lightness, α1, α2, α
3 = 0.6, β1, β2, β3 = 0, γ1, γ2, γ3 = 0.
4 and when emphasizing the change in hue, α1, α2,
α3 = 0, β1, β2, β3 = 0.6, γ1, γ2, γ3 =
It was set to 0.4.

The masks and algorithms necessary for extracting edges (rising edges, falling edges), gradients, and straight lines (ridge lines, valley lines) in 16 directions will be described below. As shown in FIG. 20, numbers are assigned from 1 to 16 in the counterclockwise direction from the top in 16 directions. Here, a mask for detecting the direction 1 is shown in FIG. Direction 5, Direction 9, Direction 13
The mask that detects
It is rotated 180 degrees and 270 degrees. Also,
A mask for detecting the direction 15 is shown in FIG. The mask for detecting the directions 3, 7, and 11 is rotated 90 degrees, 180 degrees, and 270 degrees counterclockwise. Further, a mask for detecting the direction 16 is shown in FIG.
The mask for detecting the directions 4, 8 and 12 is rotated 90 degrees counterclockwise, 180 degrees and 270 degrees. Furthermore, a mask for detecting the direction 14 is shown in FIG.
It shows in (d). The mask for detecting the directions 2, 6, and 10 is rotated 90 degrees counterclockwise, 180 degrees, and 270 degrees.

The above direction detection masks are sequentially applied while scanning the target image area pixel by pixel. Specifically, it is divided into the following two processing steps.

Step 1 Weights shown in the corresponding mask (1 or -1,0)
Is multiplied by the corresponding pixel in the image area where the windows overlap, and the absolute value of the sum is obtained. Do this for all windows and then compare their values to find the direction with the largest value. More specifically, FIG.
As shown in FIG. 2, it is assumed that pixels x and y are being processed. At this time, for example, consider a window W (i, j) of 9 × 9 pixels centered on x, y. Further, masks M1 (i, j) ... M16 (i, j) for detecting the 16 directions in FIGS. 21 (a) to 21 (d) corresponding to the window W (i, j) and the mask M1 (i, j). ) -M16 (i, j) "1"
A mask M17 for detecting the reverse inclination in which "-1" is exchanged
(I, j) ... M32 (i, j) is obtained, and each mask Mk
Synergistic sum PK of (i, j) and window W (i, j)
(K = 1 to 32) is calculated.

If this absolute value is larger than the lower limit value S1 which is distinguished from the predetermined noise, it is considered that the image area has a characteristic, and the direction and the positive / negative of the inclination (the mask is multiplied by the corresponding pixel of the image area). Make a note of the positive and negative values of the sum.

Step 2 Next, a value S2 obtained by multiplying the absolute value of the maximum inclination by 1/2 the value of the inclination 180 degrees opposite to the direction having the maximum inclination.
And the type of feature of the area is determined. First, consider the case where the inclination in the direction extracted in step 1 is positive. If the inclination 180 degrees opposite to the extracted direction is negative and its absolute value is larger than S2, this feature is "gradient".
If the absolute value of the inclination of 180 degrees opposite to the extracted direction is smaller than S2, this feature is
If the inclination in the direction opposite to the extracted direction by 180 degrees is also positive and its absolute value is larger than S2, this feature is classified as a "valley line". Also, consider the case where the inclination in the direction extracted in step 1 is negative. If the inclination in the opposite direction of 180 degrees to the extracted direction is positive and its absolute value is larger than S2, this feature is classified as "gradient", and the absolute value of the inclination in the opposite direction of 180 degrees to the extracted direction is smaller than S2. For example, this feature is classified as "falling", and if the inclination in the direction opposite to the extracted direction by 180 degrees is negative and its absolute value is larger than S2, this feature is classified as "ridgeline". Through the above procedure, five types of features are extracted. Based on this result, the process shifts to the mark recognition process similar to that shown in the first embodiment.

[Embodiment 3] In this embodiment, not only is the mark judged, but the code previously given to the mark is also deciphered. In this embodiment, a small area used for recognizing a mark is divided into eight areas as shown in FIG. 23. For example, four areas 61 on the inner side recognize the mark, and four areas 62 on the outer side mark the mark. It is to decipher the code. Of course, the relationship between the inside and the outside may be reversed.

In this embodiment, the four inner regions 61
Regarding the above, the mark is discriminated by the same method as in the first and second embodiments. On the other hand, the process of feature extraction is the same for the four outer regions 62. And the four outer regions 6
If various information is given in advance to the feature ratio extracted for 2, it is possible to read the information by using the mark determination processing method as it is. Of course, it is also possible to use a generally known cryptographic barcode other than the feature ratio.

The information that is conveniently stored as a cipher is the type of the manuscript read by the scanner. By applying this cipher, the arrangement angle when the manuscript is placed on the scanner can be determined. It is also possible to know. For example, as shown in FIG. 23, the 1/4 arc Ma in the encrypted part is changed from the color of the other part of the mark. When the mark portion is recognized by being divided into four small regions (R1 to R4), as shown in FIG. 24, the small regions (R1 to R4) are divided into quarters.
Since the arc Ma extends, it is possible to determine the orientation of the document 70 from the ratio of the total number of features represented by the colors used for the quarter arc Ma. In FIG. 24, since the marks Ma of this color are present at the same ratio in the two right small areas (R2, R4), it can be seen that the marks are inclined by 45 degrees. This information is the position of the document 70 itself being printed marks from being grasped, One coating when the mark system has found
It is possible to perform processing such as dusting . However, in order to perform such processing only on the printed matter with the mark read by the scanner, the system needs to store the size of the target printed matter, the position of the mark, and the direction of the mark in advance. .

Example 4 Unlike the first and second examples, the feature extraction unit 22 according to this example determines the direction component intersecting at right angles with the direction in which the ink on which the mark is printed is fading in the color space. The mark recognition processing unit 24 (1) compares the calculated feature ratio with the previously obtained feature ratio of the mark to determine whether they match or are similar to each other. It is to determine whether or not. According to this embodiment, the sensitivity of the color gamut change in the fading direction of the ink can be made dull, so that the mark can be accurately recognized even if the ink on which the mark is printed is fading. . For example, FIG. 25 is a schematic two-dimensional view showing the color distribution of the ink of the mark including the new printed matter and the old printed matter with the mark.
It shows in (a). In the figure, the two closed loop curves show the color distribution areas of ink A and ink B, and the hatched area shows the old ink distribution area. Then, the main component of the color distribution shown in FIG.
When the main components (first main component and second main component) are shown in (b), the second main component intersects the first main component at a right angle, and although the first main component has a large fading change, It is understood that the fading changes of the two main components are small. Therefore, if the feature ratios of the marks are compared along the second principal component, it is possible to recognize marks that are resistant to ink fading.

[0057]

As described above, according to the present invention, since the means for recognizing the mark in the image to be processed is provided, the image processing for the image to be processed is performed depending on whether the mark is recognized or not. Can be made different, and the abuse of the reproduced image can be effectively prevented according to the recognition of the mark.

In particular, according to the second aspect of the invention, since the mark position information storage means is added, the mark position information can be effectively used and the mark recognition processing and the mark image processing can be realized more efficiently. It is possible. According to the third aspect of the invention, since the mark image processing means that the recognized mark means is added, the image processing content that the mark means is immediately executed when the mark is recognized, and for example, reproduction is not possible. It is possible to realize a measure such as stopping the output of the reproduced image for the processing target image. Furthermore, according to the invention of claim 4, the mark position information from the mark position storing means is used to perform the specific mark image processing on the mark part. Therefore, for example, the mark part is painted black. On the contrary, it is possible to implement a measure such that the image to be processed and the reproduced image are made different at the mark portion, such as whitening out.

According to the invention of claim 5, in the type in which the search window is divided into smaller than the size of the mark, at the time when it is determined that the image feature pattern in the search window is not part of the mark, Since the determination process for the mark is completed and the process proceeds to the next process, the mark determination process can be performed quickly.

Further, according to the invention of claim 6, by devising a part of the mark, it is possible to easily determine the layout relationship of the document.

Further, according to the invention described in claim 7, since the characteristic information used in the mark judging means is extracted by using the color coordinate system which is insensitive to the fading change direction of the mark, the mark can be determined. Even if the color changes, the mark can always be recognized accurately.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a configuration of an image processing apparatus according to the present invention.

FIG. 2 is an explanatory diagram illustrating a configuration of a color image forming apparatus according to a first embodiment.

FIG. 3 is a block diagram illustrating details of an image processing unit according to the first embodiment.

FIG. 4 is a block diagram showing details of a characteristic pixel extraction unit according to the first embodiment.

FIG. 5 is a flowchart showing an outline of an image processing process according to the first embodiment.

FIG. 6 is an explanatory diagram illustrating a usage example of a feature buffer according to the first embodiment.

FIG. 7 is an explanatory diagram showing a relationship between a target pixel and surrounding pixels in the first embodiment.

8A to 8C are explanatory diagrams showing examples of characteristic pixels extracted in the first embodiment.

FIG. 9 is an explanatory diagram showing an example of a correspondence relationship between image shades and characteristic pixels in the first embodiment.

FIG. 10 is an explanatory diagram showing a pixel processing order by a noise removing unit according to the first embodiment.

FIG. 11 is an explanatory diagram showing a mode in which noise is not deleted as noise by the noise deletion unit according to the first exemplary embodiment.

FIG. 12 is an explanatory diagram showing another mode in which noise is not deleted as noise by the noise deletion unit according to the first exemplary embodiment.

FIG. 13 is an explanatory diagram showing a specific example of a mark according to the first embodiment.

FIG. 14 is an explanatory diagram showing a partitioned area when a mark is recognized in the first embodiment.

FIG. 15 is an explanatory diagram illustrating an operation example of a search window according to the first embodiment.

FIG. 16 is a flowchart showing a mark recognition process according to the first embodiment.

17 is an explanatory diagram schematically showing the position of a search window in the processing flow of FIG.

FIG. 18 is a block diagram illustrating an image processing unit according to a second embodiment.

FIG. 19 is a block diagram showing details of a characteristic pixel extraction unit according to the second embodiment.

FIG. 20 is a conceptual explanatory diagram of a 16-direction detection mask used in the second embodiment.

21A to 21D are explanatory diagrams showing a typical configuration example of a mask used in the second embodiment.

FIG. 22 is an explanatory diagram showing a part of a feature extraction process according to the second embodiment.

FIG. 23 is an explanatory diagram showing a partitioned area at the time of mark recognition according to the third embodiment.

FIG. 24 is an explanatory diagram showing a mark recognition process according to the third embodiment.

25A is an explanatory diagram schematically and two-dimensionally showing the color distribution of the ink of the mark used in the mark recognition according to Example 4, and FIG. 25B is the main component of the color distribution of FIG. It is an explanatory view shown.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Image reading means, 2 ... Feature extraction means, 3 ... Feature buffer, 4 ... Mark registration means, 5 ... Mark determination means, 6 ... Search window, 7 ... Mark position storage means, 8 ... Mark image processing means, 9 ... Manuscript, M ... Mark, Ma ... Orientation determination component

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location 7459-5L G06F 15/70 330 Z H04N 1/46 Z

Claims (7)

[Claims] 1. A feature extraction means (2) for extracting image feature information pixel by pixel from input image data read by the image reading means (1), and an image extracted by this feature extraction means (2). A feature buffer (3) for writing the feature information for each pixel, and a mark registration means (4) for pre-registering the image feature information of the mark (M) having a size across a plurality of pixels and consisting of predetermined image feature information. , The image feature information in the feature buffer (3) is sequentially searched in the search window (6) divided into the same size as the mark (M) or smaller, and the image feature information pattern in the search window (6) is It is checked whether or not all or some of the image feature information patterns of the mark (M) registered in the mark registration means (4) match or resemble, and if they match or resemble, all of the mark (M). Alternatively, the image processing apparatus is provided with a mark determination means (5) for determining that the image is a part. 2. The apparatus according to claim 1, further comprising mark position storing means (7) for storing the position of the image area determined to be the mark (M) by the mark determining means (5). An image processing apparatus characterized in that. 3. The apparatus according to claim 1, further comprising the step of executing a specific mark image processing meaning of the mark (M) when the mark determining means (5) determines that the mark (M) is present. An image processing device comprising a mark image processing means (8) for performing. 4. The mark position storage means (7) according to claim 1, further comprising a mark position storage means (7) for storing the position of the image area determined to be the mark (M) by the mark determination means (5). An image processing apparatus comprising: a mark image processing means (8) for executing a specific mark image processing on a mark image area based on the mark position information of the mark position storing means (7). 5. A mark determination means (5) corresponding to a mark (M) in a type according to claim 1, wherein the search window (6) is divided into smaller pieces than the mark size. The search window (6) is sequentially moved within the image area to be searched for each divided area of the search window (6), and it is determined that the image feature information pattern in the search window (6) is not a part of the mark (M). The image processing apparatus ends the determination process in the image area corresponding to the mark (M) at the point of time, and shifts to the next process. 6. The mark (M) according to any one of claims 1 to 4, wherein the mark (M) has an arc-shaped orientation determining component (Ma) different from other mark portions existing only in at least a certain partitioned area. The mark determination means (5) recognizes at which position the orientation determination component (Ma) exists, and also determines the tilt angle of the document (9) read by the image reading means (1). An image processing device characterized by the above. 7. The mark extracting means (5) according to any one of claims 1 to 4, wherein the feature extracting means (2) uses a direction component having a low sensitivity to a fading change direction of the mark (M). An image processing device characterized by extracting feature information used in.