JP2008005177A - Image processor, image processing method, and image processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor etc., capable of improving recognition precision of a specific pattern by excluding characters from objects to be recognized without adding one-bit information indicating a character. <P>SOLUTION: The image processor includes a character area decision means of deciding a character area from input image data, a pattern recognizing means 204 of recognizing the specific pattern from the input image data, and a control means 203 of performing processing for excluding an area decided as the character area from based upon a character area decision result from objects of specific pattern recognition. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a scanner suitable for preventing reading of a manuscript from which copying of banknotes, securities, confidential documents, etc. is prohibited, a printer for printing image data read by the scanner, or an electrophotographic system The present invention relates to an image processing apparatus applied to an image forming apparatus such as a color copying machine or an MFP (Multi Function peripherals).

  In recent years, the performance of color copiers has improved remarkably, and the image quality of copied images has improved to a level that cannot be discerned with the naked eye. On the other hand, abuse of color copiers, such as copiers and scanners, has been improved. Banknotes and confidential documents are being forged using them.

  In order to prevent such forgery, it has heretofore been determined whether a counterfeit is extracted by extracting a specific pattern from an original image read by a scanner. A technique for extracting only an image in a certain color range and performing binarization processing is known.

Specifically, one of the input R, G, B color signals is binarized to generate a binarized image, and the binarized image has a shape constituting a mark in the document. A technique is disclosed in which the position information of a mark (specific pattern) is extracted by matching and the degree of conformity of whether the mark is in a predetermined arrangement is determined (for example, see Patent Document 1).
Japanese Patent Laid-Open No. 11-234515

  However, when a specific pattern is extracted from an image read by a scanner, if the image includes a character, the character may be erroneously recognized as a mark.

  On the other hand, characters can be discriminated by using the area discrimination function of the compact PDF (C-PDF).

  However, since the pattern recognition unit for recognizing a specific pattern cannot recognize that it is a character as it is, in order to recognize it as a character, separately from R, G, B (8 bits × 3 = 24 bits) data. Therefore, it is necessary to add 1-bit information for indicating a character. This increases the amount of information and increases the processing time, and also requires a new process for recognizing information indicating that it is a character.

  The present invention has been made in view of such circumstances, and without adding 1-bit information indicating the character, the character is excluded from the recognition target and the recognition accuracy of the specific pattern is improved. It is an object of the present invention to provide an image processing apparatus and an image processing method that can be achieved, and to provide an image processing program for causing a computer to execute image processing.

The above problem is solved by the following means.
(1) A character area discriminating unit that discriminates a character area from input image data, a pattern recognition unit that recognizes a specific pattern from input image data, and an area that is discriminated as a character area by the character area discrimination result is: And an image processing apparatus comprising: control means for performing processing for excluding the specific pattern recognition target.
(2) The control means extracts a color threshold for performing the specific pattern recognition, and sets the color data of a pixel determined to be a character area to a value outside the threshold range as the exclusion process. 2. The image processing apparatus according to item 1 above.
(3) Character size identification means for identifying the character size of the area determined to be the character area is provided, and the control means performs the exclusion process when the character size identified by the character size identification means is a predetermined size or less. 3. The image processing apparatus according to 1 or 2 above.
(4) The pattern recognition means compares the data of each color with a threshold range determined for each color, and recognizes a specific pattern when all the data of each color is within the threshold range. The control means performs the exclusion process by setting the data for a color with the narrowest threshold range among the threshold values of each color to a value outside the threshold range. An image processing apparatus according to 1.
(5) a step of discriminating a character region from the input image data; a step of recognizing a specific pattern from the input image data; An image processing method comprising: a step of performing a process of excluding the target to be performed.
(6) A step of extracting a color threshold value for performing the specific pattern recognition, and the color data of a pixel determined to be a character area is set to a value outside the threshold range as the exclusion process. 6. The image processing method according to 5.
(7) The image processing method according to (5) or (6), further including a step of identifying a character size of an area determined as the character area, wherein the exclusion process is performed when the identified character size is a predetermined size or less.
(8) In the pattern recognition step, each color data is compared with a threshold range determined for each color, and when each color data is all within the threshold range, a specific pattern is recognized. 8. The image processing method according to any one of the preceding items 5 to 7, wherein the exclusion process is performed by setting data for a color having the narrowest threshold range among the threshold values to a value outside the threshold range.
(9) a step of determining a character region from the input image data, a step of recognizing a specific pattern from the input image data, and a region determined to be a character region by the character region determination result is the specific pattern recognition An image processing program for causing a computer to execute a step of performing a process of excluding the target to be performed.
(10) Further causing the computer to execute a step of extracting a color threshold value for performing the specific pattern recognition, and as the exclusion process, the color data of the pixel determined to be a character area is a value outside the threshold value range. Item 10. The image processing program according to Item 9, which causes a computer to execute processing to be set in the item.
(11) The method according to (9) or (10), further causing the computer to execute a step of identifying a character size of the area determined to be the character area, and causing the exclusion process to be executed when the identified character size is a predetermined size or less. Image processing program.
(12) In the pattern recognition step, the pattern recognition step compares the data of each color with a threshold range determined for each color, and if the data of each color is all within the threshold range, Any one of the preceding items 9 to 11 in which the recognition processing is performed and the exclusion processing is performed by setting the data of the color with the narrowest threshold range among the threshold values of each color to a value outside the threshold range. The image processing program described in 1.

  According to the invention described in item (1) above, when a character area is determined from the input image data, the character area is excluded from the specific pattern recognition target of the image data. Thus, it is possible to recognize with high accuracy, and it is not necessary to add information indicating that it is a character, and the time for pattern recognition processing does not increase.

  According to the invention described in (2) above, for example, each threshold value of R, G, and B colors is extracted, and color data determined to be a character area is set outside the extracted threshold value range. Therefore, it can be excluded from binarization processing targets for specific pattern recognition with a relatively simple configuration.

  According to the invention described in the above item (3), when the character size of the area determined as the character area is equal to or smaller than the predetermined size, the character area is excluded, so that the exclusion process is performed within the minimum necessary range. Thus, erroneous recognition of a specific pattern due to the inclusion of characters can be reliably eliminated.

  According to the invention described in item (4) above, the pattern recognition means compares the color data of each color with the threshold range defined for each color, and the color data of each color is all within the threshold range. Specific patterns are recognized, and the exclusion process is performed by setting the color data of the color with the narrowest threshold range among the threshold values of each color to a value outside the threshold range. Therefore, the amount of data conversion required to set the data value outside the threshold range can be reduced, and the data value only needs to be set outside the threshold value for one color. Exclusion processing becomes easy.

  According to the invention described in item (5) above, the specific pattern recognition is not affected by characters, and can be recognized with high accuracy, and it is not necessary to add information indicating the characters, and pattern recognition is possible. Processing time does not increase.

  According to the invention described in the preceding item (6), it is possible to exclude from binarization processing targets for specific pattern recognition with a relatively simple configuration.

  According to the invention described in item (7), it is possible to reliably eliminate erroneous recognition of a specific pattern due to the inclusion of characters while performing exclusion processing within a necessary minimum range.

  According to the invention described in item (8) above, the character region exclusion process is facilitated.

  According to the invention described in item (9) above, it is possible to cause the computer to execute the character area exclusion processing.

  According to the invention described in item (10) above, it is possible to cause a computer to execute the character region exclusion process with a relatively simple configuration.

  According to the invention described in the preceding item (11), it is possible to cause the computer to execute the exclusion process within a necessary minimum range.

  According to the invention described in item (12) above, the character area exclusion process is facilitated.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing an image processing system using an image processing apparatus according to an embodiment of the present invention.

  1, the image processing system includes an image reading apparatus 100 as an image processing apparatus, an image output apparatus 101, and a personal computer (hereinafter referred to as a PC) 102. These image reading apparatus 100 and image output apparatus The apparatus 101 is connected to the PC 102 via communication lines 103 and 104, respectively.

  The image reading apparatus 100 has a function of reading an image of a document. In this embodiment, a scanner is used.

  The image output device 101 outputs image data read and image processed by the image reading device 100, and a printer is used in this embodiment.

  FIG. 2 is a block diagram showing a schematic electrical configuration of the image reading apparatus 100, the image output apparatus 101, and the PC 102.

  In FIG. 2, the image reading apparatus 100 includes a document image reading unit 201, an image processing control unit 200, a memory 205, and an image data transfer unit 206.

  The document image reading unit 201 reads an image of a document, and is composed of, for example, a CCD.

  The image processing control unit 200 includes an image processing unit 202 that performs image processing such as shading correction and interline correction on image data from the CCD 201, a CPU 203, and an image recognition unit 204 that recognizes a specific image (pattern). have.

  The image output apparatus 101 includes an image data receiving unit 209 that receives image data from the image reading apparatus 100 and an image data output unit 210 that prints the received image data on paper.

  The PC 102 includes an image data receiving unit 208 that receives image data from the image reading apparatus 100, and an image processing unit 207 that executes processing or processing instructed by the user on the image data.

  The schematic operation of the image reading apparatus 100 will be described. When an original image is read by the CCD 201, the read image data is subjected to shading correction and interline correction in the image processing unit 202 and then transferred to the image recognition unit 204. Is done. If the input image is a specific image, the image recognition unit 204 reports to the CPU 203 that the input image is a specific image. If the input image is a specific image, the CPU 203 transfers data to the PC 102 or the image output apparatus 101. To prevent illegal copying.

  FIG. 3 is a block diagram showing a specific configuration of the image processing control unit 200 in the image reading apparatus 100.

  As shown in FIG. 3, the image processing control unit 200 includes an image preprocessing unit 52, an area determination unit 53, an attribute determination unit 54, a compression unit 55, a PDF file creation unit 56, an image recognition unit 204, as main components. A CPU 203, a RAM 60, and a ROM 61 are provided, which are connected via the bus 51 and can exchange data.

  The image preprocessing unit 52 performs preprocessing on the image data from the CCD 201, and sends the preprocessed data to the region determination unit 53.

  The area discriminating section 53 executes an area discriminating process for discriminating a character area and a photograph area based on image data from the image preprocessing section 52. In this embodiment, a compact PDF (C-PDF) function is used. Is adopted.

  The attribute discriminating unit 54 extracts, for the discriminated area, a characteristic amount extraction process for extracting the characteristics of the area (number of colors used, character size, ruled line, halftone dot ratio, etc.), and the extracted characteristics The attribute information creation process for creating the attribute information of the input image (the degree of fit indicating the probability that the input image belongs to which kind of image type such as document, table, catalog, etc.) is executed.

  The compression unit 55 compresses a character image in the character region determined by the region determination unit 53, a background image other than the character image, and the like using JPEG.

  The PDF file creation unit 56 converts the compressed data into a known PDF format, and adds the attribute information to the converted file.

  The image recognition unit 204 has a pattern matching function for recognizing a specific pattern from input image data.

  The CPU 203 operates in accordance with a program stored in the ROM 61 to uniformly control the entire operation of the image reading apparatus 100 or instructs the image preprocessing unit 52, the area determination unit 53, etc. A series of processing (PDF conversion processing) from PDF conversion of the input image from the image to output to the memory 205 is executed.

  The CPU 203 performs processing for excluding the area determined as the character area from the result of the character area determination from the target for the specific pattern recognition. Specifically, the image recognition unit 204 extracts a color threshold value for performing specific pattern recognition, and as the exclusion process, the color data of a pixel determined as a character area is a value outside the threshold value range. Set to. In that case, it is preferable to forcibly exclude the color (for example, B color) having the narrowest threshold range among the three colors R (red), G (green), and B (blue) as color data.

  The ROM 61 stores a program for executing PDF conversion processing and the like. In addition, a basic feature value table showing the magnitude relationship of feature values for various images such as documents, tables, and catalogs is also stored.

  The RAM 60 serves as a work area when the CPU 203 executes various programs.

  FIG. 4 is a diagram showing an outline of a processing flow until the PDF file is stored in the memory 205.

  In FIG. 4, in step S <b> 1, image data read by the CCD 201 is transferred to the image processing unit 201.

  In step S2, preprocessing for region determination is performed on the read image data. This pre-process is a process for determining the area of the input image with higher accuracy, and is specifically a known process such as resolution conversion (from high to low) and background removal.

  In step S3, the region of the image such as a character and a photograph is determined. Then, the feature amount of the discriminated area is extracted, and attribute information is created based on the extracted feature amount. The created attribute information is also sent to the image recognition unit 204.

  In step S4, a character image is extracted from the input image based on the determination result, and the extracted character image is compressed by MMR.

  In step S5, a background image is extracted from the input image based on the determination result, and the extracted background image is compressed by JPEG.

  In step S6, the compressed character image and the background image are converted into PDF and combined. At that time, the attribute information created in step S3 is added and output as a PDF file.

  In step S 7, the output PDF file is stored in the memory 205.

  FIG. 5 is a flowchart showing the PDF conversion process.

  In FIG. 5, when the CPU 203 receives a PDF conversion instruction from the user, the CPU 203 executes a scan image acquisition process via the CCD 201 in step S10.

  The acquired image data is stored in a predetermined storage area in the RAM 60 as an input image as it is. Hereinafter, in each process, when using R (red), G (green), and B (blue) pixel data of an image, the data stored in the RAM 60 is read out.

  Next, an area determination process is executed in step S20. Specifically, the CPU 203 first performs the predetermined preprocessing on the acquired image data via the image preprocessing unit 52, and then converts the preprocessed input image to the preprocessed input image via the region determination unit 53. Based on the area discrimination processing.

  In step S <b> 30, feature amount extraction processing is executed via the attribute determination unit 104.

  In step S40, attribute information creation processing is executed. In step S50, image compression processing is executed. In step S60, PDF file creation is executed.

  FIG. 6 is a flowchart showing a subroutine of region discrimination processing (step S20).

  Steps S201 to S204 as photo region discrimination and steps S205 to S213 as character region discrimination are executed.

  In FIG. 6, in step S <b> 201, brightness calculation processing is first performed in the photographic area determination. Specifically, for example, if the input image is color image data, the pixel data of R, G, B color (255 gradations) for the pixels of the image for one page of the document is converted to brightness data using a known conversion formula. Convert to

  In step S202, a smoothing process is performed to remove noise included in the input image by applying a known smoothing filter to the generated brightness data. In step S203, a photographic block is detected based on the data subjected to the smoothing process. Here, binarization is performed on the noise-removed data using a predetermined threshold value.

  Then, labeling is performed by connecting pixels to the binarized image, and it is determined whether each labeled image is a photographic region. For example, the size of a block is larger than a predetermined size (a size corresponding to a character), and characteristics such as brightness, saturation, hue, etc. of each image in the block, for example, halftone pixels are a predetermined ratio or more A photo area is identified when a predetermined condition is indicated, such as when it is included. A block that does not satisfy the predetermined condition is not regarded as a photo area.

  In addition, the area | region discriminate | determined as a photograph area | region here contains the area | region of the image which has gradation property, such as a picture and a chart, besides a photograph. Further, the input image data stored in the RAM 60 is referred to for the saturation and hue of each pixel.

  In step S204, for each block determined to be a photo area, coordinate data of a rectangular area into which the block (photo area) is taken is stored in the RAM 60.

  On the other hand, in character region discrimination, in step S205, first, brightness calculation processing for converting the input image into brightness data is performed. This process is basically the same as the process of step S201, except that a parameter for character discrimination is used in the lightness data conversion formula.

  In step S206, smoothing processing is performed to remove noise by applying a smoothing filter to the brightness data. This smoothing is basically the same as the processing in step S202, except that a filter suitable for character discrimination is used.

  In step S207, a binarization process for generating a binary edge image by applying a known edge filter to the data subjected to the smoothing process is performed.

  In step S208, processing for detecting a ruled line included in the generated binary edge image is performed. Specifically, a ruled line included in the edge image is detected by applying a known filter for ruled line detection. In step S209, the amount of the ruled line detected (here, B / A, where A is the total number of pixels of the preprocessed input image and B is the number of pixels of the ruled line image), and in step S210, The calculated data is stored in the RAM 60, and the detected ruled line is deleted. This ruled line deletion is for improving the character discrimination accuracy.

  In step S211, character block detection processing is performed in which binary edge images from which ruled lines have been deleted are connected within a predetermined range in the main scanning direction and the multi-scanning direction, and a rectangular area surrounding the connected blocks is set. I do.

  In step S212, character determination processing is performed to determine whether or not the set character block image is a character. For example, the local shape of an image in a character block is extracted as a feature value, and whether or not the image is attributed to a character is determined from the extracted feature value. More specifically, as the local shape of the image, the curve amount, the inclination direction, the number of closed loops, the number of crosses, the number of T-shaped intersections, and the like are extracted, and the extracted feature points hold the character discrimination previously held. If the number matching with the feature points of the pattern for the above is greater than or equal to a predetermined value (threshold), it is determined as a character, and if it is less than the predetermined value, it is determined that the character is not a character.

  For character discrimination, other methods such as pattern recognition based on a dictionary for character discrimination can be used.

  In step S213, the coordinate data of the block (character area) determined to be a character is stored in the RAM 60.

  In step S214, a process for detecting a halftone dot area included in the character area is executed. This halftone dot detection is performed, for example, by applying a known halftone dot detection filter to pixels in a character area, and when the number of isolated points in a predetermined range exceeds a predetermined number, the range is set as a halftone dot area. .

  In step S215, the amount of the halftone dot (here, D / C, where C is the total number of pixels in the character area and D is the number of pixels in the halftone area) is calculated from the detected halftone area. In S216, after the calculated data is stored in the RAM 60, the point images scattered in the halftone dot region are removed.

  Subsequently, in step S217, a process of binarizing each pixel in the character area from which halftone dots have been removed using a predetermined threshold to separate a character image in the character area from a portion other than the character (background). Do.

  In step S218, the data of the coordinates of each pixel constituting the character image separated by binarization (hereinafter referred to as “character map”) and the data of the coordinates of the background portion are stored in the RAM 60. In step S219, the number of character colors is calculated, and then the process returns to the main routine.

  In calculating the number of character colors, for example, pixels constituting a character are specified from a character map, and R, G, B color pixel data is converted into data representing hue for each specified pixel, By comparing with the threshold value, find out which color among the predetermined colors (black, blue, red, yellow, white, etc.) belongs to that number (number of black pixels, blue pixels, etc.) This is done by factoring The calculated data is stored in the RAM 60.

  Note that the method for discriminating the area of the character or the like is not limited to the above, and other known methods can be used.

  FIG. 7 is a flowchart showing the contents of a subroutine of the feature amount extraction process (step S30) of FIG.

  In FIG. 7, in step S31, a character size calculation process is performed. For calculating the character size, for example, referring to a character map, for each character image, the number of pixels arranged in the direction corresponding to the height direction of the character is counted, and the size (point) of the character is calculated from the number of pixels. This is done by detecting. Here, data indicating the largest number of points among characters is stored in the RAM 60.

  Next, in step S32, a background area amount calculation process is performed. Here, the background area is a background in the character area, and the background is other than white (for example, corresponds to the background in the character area of a newspaper).

  As a method for calculating the amount of the background area, for example, the background coordinate data stored in step S218 in FIG. 6 is referred to, and the background color is determined based on the R, G, B color pixel data of the input image. This is done by extracting the other pixels, setting the number of extracted pixels as E, calculating the value of E / C, and setting this value as the amount of the background area. The calculated data is stored in the RAM 60.

  In step S33, the number of colors of the image is calculated. This process is the same as the process for calculating the number of character colors in step S219 in FIG. 6. Here, the pixels of the image determined to be a photographic region are specified, and the R, G, B colors of each specified pixel are specified. This is a process of determining which color of the plurality of colors belongs to each pixel from the pixel data and counting the number. The calculated data is stored in the RAM 60.

  After performing the positional relationship detection process in step S34, the process returns. This positional relationship detection process detects the characteristics of the input image from the relationship between the arrangement position of the character area and the arrangement position of the photo area in the input image.

  Next, a specific image pattern recognition process performed by the image recognition unit 204 of the image processing control unit 200 in the image reading apparatus 100 will be described.

  In the image recognition unit 204, basically, when R, G, B color image data is input, the image data is binarized, and then a specific pattern is generated from the input image. Recognizing and performing pattern matching processing for comparison with a predetermined pattern, and finally, by relevance processing, for example, if the recognition has a relevance of 70% or more, it is determined as “similar”.

  By the way, in this embodiment, in the binarization process by the image recognizing unit 204, each of the brightness of R, G, and B colors (0 to 255 gradations) for one pixel has a predetermined threshold. If it is within the value range (between the upper limit threshold and the lower limit threshold), the binarization process is executed as the binarization process target, and the brightness of the R, G, and B colors is set. If any of them is out of the threshold range, it is out of the binarization process target. For example, when R = 100, G = 100, and B = 100, binarization processing is performed if any of them is within the threshold range, but when R = 100, G = 100, and B = 10 If the value of B is outside the threshold range, the condition that all the colors are within the threshold range (AND condition) is not satisfied, so that the pixel is excluded from the binarization processing target. become.

  Here, if the image read by the CCD 201 includes a character, it can be determined as a character by the C-PDF region determination function. In this case, the region determined as a character region is used. In this embodiment, the CPU 203 performs the following operation so that the binarization processing by the image recognition unit 204 is not performed.

  That is, if it is determined that the character is one of the R, G, and B colors constituting the character, for example, if the brightness value of B is out of the threshold range, the binarization process is originally performed. Since it is not implemented, it is left as it is. When the value of B is within the threshold range, the values of R and G are confirmed. When at least one of R and G deviates from the respective threshold values, the state remains as it is. If both R and G are within the respective threshold ranges, the value of B is converted by the following equation.

B threshold (upper limit threshold) -read B value + 1 = B1
Read B value + B1 = B2 (B value after conversion)
Thus, by converting the difference between the B threshold value (upper limit threshold value) and the read B value to a value obtained by adding “1”, the converted B value is the B threshold value ( The upper limit threshold) is exceeded by 1, and it is out of the threshold range. As a result, even if the image data is sent to the image recognition unit 204, it is automatically excluded from the binarization processing target for the specific image recognition by the image recognition unit 204.

  Further, the value of B may be uniformly converted to “255” without using the above conversion formula.

  In any case, if one of R, G, and B is outside the threshold range, it can be excluded from the target image for binarization processing. Therefore, when B is changed, R, It is not necessary to perform color conversion for G.

  Similarly, when R is changed, it is not necessary to perform color conversion for G and B, and when G is changed, it is not necessary to perform color conversion for R and B.

  As described above, when the character area is determined from the input image data, the character area is excluded from the specific pattern recognition target of the image data, so that the specific pattern recognition process is not affected by the character. The pattern recognition process for the specific image can be performed with high accuracy, and the recognition process is performed except for the character area, so that the time for the pattern recognition process is shortened.

  Moreover, the image recognition unit 204 is compared with a case where 1-bit information is added to indicate that the character is a character, and the image recognition unit 204 performs an operation of removing from the binarization processing target based on the added information. It is not necessary to provide a special function, and the existing image recognition unit 204 can be used as it is, which is very convenient.

  It is sufficient to set any one value of R, G, and B outside the threshold range, and any value of R, G, and B may be set. For the color with the narrowest range (generally B), the value should be set to a value outside the threshold range. This is because the amount of data conversion required to set the threshold value outside is small and processing is easy.

  FIG. 8 is a flowchart showing processing for excluding a character region from binarization processing. This process is also executed by the CPU 203 operating in accordance with a program stored in the TOM 61.

  In FIG. 8, in step S60, image data composed of each color of R, G, and B is acquired. In step S61, whether or not all the pixels are character determination processing and whether or not exclusion processing for characters is complete. If all the pixels are not completed (NO in step S61), in step S62, characters are determined for each pixel for each of R, G, and B, and the process proceeds to step S63.

  In step S63, it is determined whether or not the pixel is a character. If it is not a character (NO in step S63), the process returns to step S61 because it is a binarization target. If it is a character (YES in step S63), it is not subject to binarization, so in step S64 it is confirmed whether the value of B is within the threshold range, and the process proceeds to step S65.

  In step S65, it is determined whether or not the value of B is within the threshold range and satisfies the binarization condition. If it is out of the threshold range and the binarization condition is not satisfied (NO in step S65). ) Since it is originally out of the binarization target, the process returns to step S61 in order to process the next pixel. If the value of B is within the threshold range and satisfies the binarization condition (YES in step S65), in step S66, R and G are checked in order, and then the process proceeds to step S67.

  In step S67, it is determined whether or not the values of R and G satisfy the binarization condition within the threshold range. If at least one of R and G is outside the threshold range and does not satisfy the binarization condition (NO in step S67), the process returns to step S61 to process the next pixel, and the values of R and G are If both satisfy the binarization condition within the threshold range (YES in step S67), the process proceeds to step S68.

  In step S68, when this pixel is sent to the image recognition unit 204, the binarization process is performed, so the value of B is changed. The value of B is changed as follows. That is, for example, assuming that the upper limit threshold value of B = 120 and the read B = 100, B1 = 120−100 is calculated, and B = 100 + B1 (20) + 1 = 121.

  Thus, by forcibly setting the value of B outside the threshold value range of 121, the character is excluded from the binarization target. Then, in order to perform processing for the next pixel, the process returns to step S61.

  If the processing is completed for all pixels (YES in step S61), in step S69, the image recognition unit 204 performs binarization processing for pattern recognition of the specific image.

  By the way, as described in step S31 of FIG. 7, in this embodiment, the character size is identified when it is determined to be a character area. However, by using the identified character size, When the font size is a predetermined size, for example, 20 points or less, the exclusion process may be performed.

  In other words, the specific image recognition processing by the image recognition unit 204 is generally performed by determining the relationship between the pattern matching with the basic pattern and the arrangement, but if the font size exceeds, for example, 20 points, it is specified by pattern matching. Even if it is recognized as an image, it tends to be determined that it is not a specific image due to the arrangement. For this reason, characters with a size exceeding 20 points are not erroneously recognized even if they are binarized as they are.

  On the other hand, for example, when the number of points is 20 points or less, it is easy to recognize not only the pattern matching but also the arrangement relationship. In this way, by performing the exclusion process only when the size is equal to or smaller than a certain size, the number of exclusion processes is reduced, so that the necessary minimum process is required and the process becomes easy.

  FIG. 9 is a flowchart showing processing for excluding a character region from binarization processing in consideration of such character font size. This process is also executed by the CPU 203 operating in accordance with a program stored in the TOM 61.

  The same steps as those in FIG. 8 are denoted by the same step numbers and the description thereof is omitted.

  In FIG. 9, processing for determining the font size of the character in step S631 is added after the processing for determining whether or not the character exists in step S63.

  That is, if the pixel is a character in step S63 (YES in step S63), it is determined in step S631 whether or not the font size of the character is less than a predetermined point, for example, 20 points or less. If it exceeds (NO in step S631), there is no problem even if binarization is performed, and the process returns to step S61. If the font size of the character is 20 points or less (YES in step S631), since erroneous recognition occurs, it is necessary to exclude this from the binarization target, and the process proceeds to step S64. The following processing procedure is the same as that in FIG.

  In the embodiment described above, a scanner has been described as an example of the image processing apparatus 100. However, an image forming apparatus such as a color copying machine, a personal computer, or the like may be used as the image processing apparatus.

1 is a configuration diagram illustrating an image processing system in which an image processing apparatus according to an embodiment of the present invention is used. It is a schematic block diagram which similarly shows the electric constitution of an image processing system. It is a block diagram which shows the specific structure of the scanner in an image processing system. It is a sequence which shows a process until a PDF file is stored. It is a flowchart which shows a PDF conversion process. It is a flowchart which shows the subroutine of the area | region discrimination | determination process (step S20) of FIG. It is a flowchart which shows the subroutine of the feature-value extraction process (step S30) of FIG. It is a flowchart which shows the process which excludes a character area from the object of a binarization process. It is a flowchart which shows the process which excludes a character area from the object of a binarization process in consideration of the font size of a character.

Explanation of symbols

53 Region discriminating unit 54 Attribute discriminating unit 100 Image reading device (image processing device)
200 Image Processing Control Unit 201 Document Reading Unit 202 Image Processing Unit 203 CPU
204 Image recognition unit (pattern recognition means)

Claims (12)

Character area discrimination means for discriminating a character area from input image data;
Pattern recognition means for recognizing a specific pattern from input image data;
Control means for performing processing for excluding the area determined as the character area from the character area determination result from the target for performing the specific pattern recognition;
An image processing apparatus comprising:   The control means extracts a color threshold value for performing the specific pattern recognition, and sets the color data of a pixel determined as a character area to a value outside the threshold value range as the exclusion process. The image processing apparatus according to 1. Character size identifying means for identifying the character size of the area determined as the character area,
The image processing apparatus according to claim 1, wherein the control unit performs the exclusion process when the character size identified by the character size identification unit is a predetermined size or less. The pattern recognition means compares the data of each color with a threshold range determined for each color, and recognizes a specific pattern when the data of each color is all within the threshold range,
The said control means performs the said exclusion process by setting the data to the value outside a threshold value range about the color with the narrowest threshold value range among the threshold values of each color. An image processing apparatus according to 1. Determining the character area from the input image data;
Recognizing a specific pattern from the input image data;
Performing a process of excluding the area determined as the character area from the character area determination result from the target for performing the specific pattern recognition; and
An image processing method comprising: Extracting a color threshold for performing the specific pattern recognition,
The image processing method according to claim 5, wherein as the exclusion process, color data of a pixel determined to be a character area is set to a value outside the threshold range. Identifying the character size of the area determined to be the character area,
The image processing method according to claim 5, wherein the exclusion process is performed when the identified character size is equal to or smaller than a predetermined size. In the pattern recognition step, each color data is compared with a threshold range determined for each color, and when each color data is all within the threshold range, a specific pattern is recognized,
The image processing according to any one of claims 5 to 7, wherein the exclusion processing is performed by setting data of a color having a narrowest threshold range among the threshold values of each color to a value outside the threshold range. Method. Determining the character area from the input image data;
Recognizing a specific pattern from the input image data;
Performing a process of excluding the area determined as the character area from the character area determination result from the target for performing the specific pattern recognition; and
An image processing program for causing a computer to execute. Further causing the computer to execute a step of extracting a color threshold for performing the specific pattern recognition,
10. The image processing program according to claim 9, wherein the image processing program causes the computer to execute a process of setting color data of a pixel determined to be a character area to a value outside the threshold range as the exclusion process. Further causing the computer to execute a step of identifying a character size of the area determined to be the character area;
The image processing program according to claim 9 or 10, wherein the exclusion process is executed when the identified character size is equal to or smaller than a predetermined size. In the pattern recognition step, the pattern recognition step compares each color data with a threshold range determined for each color, and recognizes a specific pattern when all the color data is within the threshold range. Let
The image according to any one of claims 9 to 11, wherein the exclusion process is performed by setting the data of a color having the narrowest threshold range among the threshold values of each color to a value outside the threshold range. Processing program.

Images