JP5146199B2 - Difference extraction apparatus and difference extraction program - Google Patents

Description

  The present invention relates to a difference extraction apparatus and a difference extraction program.

  There are many cases in which a paper document is manually added, deleted, or deleted using a correction liquid, or the document data is edited using a computer or the like. To make any changes to the original document, such as handwritten additions or seals to these original documents, additions using computers, deletions, etc. There are many needs to extract a difference portion by comparing a document before correction with a document after correction. Conventionally, a technique has been proposed in which image data to be compared is divided into regions and a difference is extracted by comparing the divided images (for example, Patent Documents 1 and 2).

JP 2004-178002 A JP 2004-287682 A

  By the way, when performing difference extraction, if a comparison target image includes a specific type of image such as a photographic image or a pattern image, these images may change in density or color due to the influence of scanner characteristic fluctuations. Since it is likely to occur, the possibility that a difference is extracted increases.

  An object of the present invention is to reduce the possibility that a difference between images is extracted even if a document is not modified.

The difference extraction apparatus according to the present invention includes a receiving unit that receives a first image and a second image that is a difference extraction target between the first image and at least one of the first and second images. The invalid area specifying means for specifying an area including an image for which an effective processing result cannot be obtained when executing the difference extraction process as an invalid area, and the invalid area is excluded from each of the first and second images Difference extraction processing means for performing difference extraction processing by comparing image areas, and a differential extraction image obtained as a result of the processing by the difference extraction processing means in a manner capable of discriminating the invalid area specified by the invalid area specifying means have a, and an image output means for outputting said image output unit, wherein when printing the difference extraction image, the image of the invalid area portion identified by the invalid area specifying means from the first and second of each image Turn off Out, characterized by printing separately from each image the difference extraction image in the clipped invalid area portion.

The difference extraction processing means includes a feature point extraction unit that extracts a feature point from an image area excluding the invalid area of each of the first and second images, and an invalid area of each of the first and second images. A correction unit that corrects the position and distortion of at least one of the first and second images by associating feature points extracted from the excluded image region, and the first and second corrected after the correction unit A difference extraction unit that performs difference extraction by comparing each image, and the invalid area specifying unit is further included in the image from at least one of the first and second images. A type of the partial image and a region of the partial image, and the feature point extraction unit includes the specified partial image in the image region excluding the invalid region of at least one of the first or second images. Depending on the type of And extracting feature points from the region of the partial image.

Difference extraction program according to the present invention, a computer, a first image, a second receiving means for receiving an image comprising the difference extraction target from the first image, the first or the second image at least one of An invalid area specifying means for specifying, as an invalid area, an area including an image from which an effective processing result cannot be obtained when executing difference extraction processing, and removing the invalid area from the first and second images. Difference extraction processing means for performing difference extraction processing by comparing image areas, and a difference extraction image obtained as a result of the processing by the difference extraction processing means in a manner capable of discriminating the invalid area specified by the invalid area specifying means image output means for outputting, to function as the image output unit, the case of printing a difference extraction image, no specified by the invalid area specifying means from the first and second of each image Cut out the image of the area portion, wherein the print separately from each image the difference extraction image in the clipped invalid area portion.

According to the first and third aspects of the invention, it is possible to reduce the possibility that a difference between images is extracted even if the original is not corrected.

Further , according to the first and third aspects of the present invention, it is possible to visually confirm whether or not there is a difference in the invalid area that is not used in the difference extraction process.

According to the second aspect of the present invention, it is possible to reduce a possibility that a difference between images is extracted as compared with a case where correction of the position and distortion of an image is not performed using a feature point corresponding to the type of image. Can do.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.

Embodiment 1 FIG.
FIG. 1 is a block configuration diagram showing an embodiment of a difference extraction apparatus according to the present invention, and FIG. 2 is a hardware configuration diagram of a computer forming the difference extraction apparatus in the present embodiment. The computer forming the difference extraction device in the present embodiment can be realized with a general-purpose hardware configuration that has existed in the past. That is, as shown in FIG. 2, the computer includes a CPU 1, a ROM 2, a RAM 3, an HDD controller 5 to which a hard disk drive (HDD) 4 is connected, a mouse 6 and a keyboard 7 provided as input means, and a display 8 serving as an image output destination. And an input / output controller 11 for connecting a printer 9 and a scanner 10 for reading an original document, and a network controller 12 provided as a communication means to an internal bus 13.

  Returning to FIG. 1, the difference extracting apparatus according to the present embodiment includes an image reading unit 20, an invalid area specifying unit 30, a difference extracting unit 40, and an image output unit 50. In general, the difference extraction apparatus, when an original (hereinafter referred to as “corrected original”) is generated by correcting the original original (hereinafter referred to as “original before correction”), the corrected original. The image reading unit 20 uses the scanner 10 to read the pre-correction original and the post-correction original to be compared in order to extract the difference. A pre-correction image and a post-correction image are generated. The invalid area specifying unit 30 extracts an area of an image type set in advance as an image from which an effective processing result cannot be obtained as a result of the difference extraction process from at least one of the pre-correction image and the post-correction image. Then, the extracted image area is specified as an invalid area. The difference extraction unit 40 generates a difference extraction image indicating a difference between the pre-correction image and the post-correction image by performing the difference extraction image processing.

  The difference extraction unit 40 includes an overall position / distortion correction unit 41, an approximate difference extraction unit 42, a divided region calculation unit 43, a region division unit 44, a local position / distortion correction unit 45, a detailed difference extraction unit 46, and a difference synthesis unit 47. Although it has, these components 41-47 are demonstrated with description of operation | movement.

  The image output unit 50 outputs the difference extraction image obtained by the difference extraction unit 40. In particular, the image output unit 50 according to the present embodiment outputs the difference extracted image so that the invalid area specified by the invalid area specifying unit 30 can be identified. The area incorporation unit 51 incorporates the invalid area part into the difference extraction image in a state where it can be distinguished from the part other than the invalid area.

  FIG. 3 is a block diagram of the overall position / distortion correction unit 41 shown in FIG. The overall position / distortion correction unit 41 in the present embodiment includes a first feature point extraction unit 411, a second feature point extraction unit 412, a corresponding feature point determination unit 413, a correction coefficient calculation unit 414, and a position / distortion correction unit 415. Have. These components 411 to 415 will be described together with the description of the operation.

  Each component 20-50 in a difference extraction apparatus is implement | achieved by the cooperative operation | movement of the program which operate | moves with the computer which forms the difference extraction apparatus, and CPU1 mounted in the computer.

  Further, the program used in this embodiment can be provided not only by communication means but also by storing it in a computer-readable recording medium such as a CD-ROM or DVD-ROM. The program provided from the communication means or the recording medium is installed in the computer, and various processes are realized by the CPU of the computer sequentially executing the installation program.

  Next, the difference extraction process in the present embodiment will be described with reference to the flowchart shown in FIG.

  In order to extract the correction portion added to the document, the image reading unit 20 scans and reads the uncorrected document to be compared and the corrected document to be read by the scanner 10. “Pre-correction image”) and a read image of the original after correction (hereinafter “corrected image”) are generated (step 101).

  Each image of the pre-correction image and the post-correction image is input to the invalid area specifying unit 30 and the difference extracting unit 40, respectively. In FIG. 1, both the uncorrected image and the corrected image are input to the invalid area specifying unit 30, but actually, the uncorrected image or the corrected image as will be apparent from the following description. It is sufficient that at least one of the rear images is input.

  When the sent pre-correction image or post-correction image is received, the invalid area specifying unit 30 specifies the invalid area by processing the received image as follows (step 102). Here, the invalid area is an image included in the difference extraction target image (pre-correction image or post-correction image) and is considered to be an image in which an effective processing result cannot be obtained as a result of the differential extraction processing. Is a region that contains For example, photographic images, images, graphics, and gradation images (hereinafter collectively referred to as “image images”) are affected by variations in the characteristics of the scanner, so that the black and white pixels after binarization are actually The possibility of a large difference is high. Feature points are extracted from each image as a pre-process for aligning the images to be compared. However, if the pixel values of each image differ greatly, the feature points extracted in this processing will differ greatly. Therefore, there is a possibility that the accuracy of the positional distortion correction is lowered.

  As described above, if the difference extraction process is performed on a type of image for which high-precision difference extraction cannot be expected, various adverse effects such as a reduction in the processing accuracy described above are likely to be caused. Examples of the types of images that are likely to cause such harmful effects include the above-described photographic images and design images, and images that include small point characters. Therefore, in the present embodiment, an area including an image that is not suitable for the difference extraction among the difference extraction target images is specified as an invalid area in order to exclude the difference extraction process.

  Since the invalid area is as described above, in step 102 in FIG. 4, when the invalid area specifying unit 30 accepts the image before correction or the image after correction, the invalid area is received by performing image processing on the image. The type of image included in the selected image is determined. Then, when an area including an image of a preset type as an image type to be an invalid area such as a photographic image or a pattern image is extracted, the area is specified as an invalid area. Note that, among the image areas included in the difference extraction target image, an area on which the difference extraction process is performed, in other words, an area other than the invalid area included in the image is referred to as an “effective area”.

  As described above, the invalid area specifying unit 30 specifies either the pre-correction image or the post-correction image as the processing target in order to specify the non-processing area in the difference extraction target image as described above. And it is sufficient. When the invalid area is identified in this way, the invalid area identifying unit 30 uses the position information where the invalid area appears in the image, specifically the coordinate information on the image, as the overall position / distortion correcting unit 41 of the difference extracting unit 40. And roughly to the difference extraction unit 42 and further to the image output unit 50.

  In addition, as a wrong result of the difference extraction process, there are a case where there is no difference and a case where the difference is erroneously extracted, and a case where there is a difference and no difference is extracted because there is no difference. It is done. Since the small point character may correspond to the latter in particular, in this embodiment, the image region including the small point character is also included in the invalid region in order to improve the reliability of the processing result.

  When the position information of the invalid area is sent, the overall position / distortion correction unit 41 in the difference extraction unit 40 determines the total position from the area (effective area) excluding the invalid area of both the pre-correction image and the post-correction image. Feature points for distortion correction are extracted (step 103). When each image is a multi-valued image such as gray or color, each image is binarized. More specifically, as shown in FIG. 3, the first feature point extraction unit 411 detects the invalid area sent from the invalid area specifying unit 30 from the pre-correction image sent from the image reading unit 20. An effective area excluding the area corresponding to the position information is extracted, and feature point extraction processing is performed in the effective area. Similarly, the second feature point extraction unit 412 also removes the effective area from the corrected image sent from the image reading unit 20 by excluding the area corresponding to the position information of the invalid area sent from the invalid area specifying unit 30. Are extracted, and feature point extraction processing is performed in the effective area. Then, the corresponding feature point determination unit 413 specifies a corresponding feature point pair from the extracted feature points of both images. Subsequently, the correction coefficient calculation unit 414 uses the feature point pair, for example, for correction of the overall position distortion for correcting the position and distortion of the post-correction image so that the position is matched with respect to the pre-correction image. Correction coefficients (movement, scaling ratio, rotation angle) are calculated. The position / distortion correction unit 415 corrects the positional distortion of the corrected image using the calculated correction coefficient (step 104).

  When the pre-correction image and the post-correction image after the positional distortion correction are received from the overall position / distortion correction unit 41, the approximate difference extraction unit 42 extracts the approximate difference for the effective area of each image (step 105). Subsequently, the divided region calculation unit 43 analyzes the approximate difference image to identify a region likely to have a difference, and the region division unit 44 determines the portion corresponding to the identified region and the surrounding region as the pre-correction image. And cut out from the corrected image corrected for positional distortion. Then, the local position / distortion correction unit 45 performs alignment between the partial image region cut out from the pre-correction image serving as the reference image for alignment and the corresponding partial image region cut out from the post-correction image. Correct the positional distortion of the corrected image. Upon receiving the pre-correction image and the post-correction image after the positional distortion correction from the local position / distortion correction unit 45, the detailed difference extraction unit 46 compares the extracted region images in detail and extracts the detailed difference. (Step 106). This detailed difference extraction process is repeated for all partial image areas.

  Note that the approximate difference extraction may be, for example, counting the number of black pixels in the effective area of each image and extracting whether there is a difference between them. In this case, when the difference in the number is outside a certain range, it may be considered that there is a possibility that there is a difference between both images.

  Then, when analyzing the difference image, for example, a histogram indicating the distribution of the appearance frequency of black pixels in the effective area of both images may be generated to specify an area likely to have a difference.

  Further, the extraction of the detailed difference is, for example, confirming the appearance position of the black pixel of each image for each partial image area where the difference is likely to occur, and the difference is determined when there are a certain number or more of black pixels at different appearance positions. It may be considered that there is.

  The process in the difference extraction unit 40 described above may be basically the same as the conventional difference extraction process, except that the process is performed on the effective area excluding the invalid area from the difference extraction target image. Finally, the difference synthesizer 47 synthesizes the images divided for the detailed difference extraction to generate one difference extracted image (step 107).

  When the image output unit 50 receives the difference extraction image sent from the difference synthesis unit 47 in addition to the position information of the invalid region sent from the invalid region specifying unit 30, the image output unit 50 incorporates the invalid region in the image output unit 50. The unit 51 incorporates the received difference extraction image so that the invalid region can be visually distinguished from the effective region and the difference extraction region in the effective region (step 108). And the image output part 50 outputs a difference extraction image (step 109). The output destination may be an external device via the printer 9, the display 8, the HDD 4, or the network controller 12.

  In addition, as a method of making it possible to determine the invalid area, a frame surrounding the invalid area on the difference extraction image is drawn, shaded, or the invalid area is colored with a different color, May be combined. Alternatively, for example, when printing a differential extraction image, an image of the invalid area portion may be cut out from each image before and after correction and printed separately from the differential extraction image. In the present embodiment, since the difference extraction process is not performed for the invalid area, the user needs to confirm the presence / absence of the difference by visual recognition. Therefore, the image of the invalid area portion cut out from each image for easy comparison. Print side by side.

Embodiment 2. FIG.
In the first embodiment, the invalid area is unconditionally excluded from the target area for the difference extraction process. In the present embodiment, feature points can be extracted by a method according to the type of image included in an effective area other than the invalid area specified in the first embodiment.

  FIG. 5 is a block diagram showing the difference extraction apparatus in the present embodiment. In addition, the same code | symbol is attached | subjected to the same component of Embodiment 1, and description is abbreviate | omitted. The present embodiment may be the same as the first embodiment except that an area separation unit 60 is provided instead of the invalid area specifying unit 30 shown in FIG.

  Next, difference extraction processing in the present embodiment will be described using the flowchart shown in FIG. The same processes as those in the first embodiment shown in FIG.

  The pre-correction image and the post-correction image generated by the image reading unit 20 reading the pre-correction original and the post-correction original are input to the region separation unit 60 and the difference extraction unit 40, respectively (step 101). Note that at least one of the pre-correction image and the post-correction image may be input to the region separation unit 60.

  Upon receiving the sent pre-correction image or post-correction image, the region separation unit 60 analyzes the received image to determine the type of image included in the image, and classifies the image by type. . That is, it is determined which type of image is formed in which area of the received image. The types of images are classified into, for example, characters, line drawings, table ruled lines, photographs, designs, and other types. In the present embodiment, a region of an image belonging to a photo, a pattern, and other types is specified as the invalid region described in the first embodiment. When the images are classified in this manner and the invalid area is specified (step 201), the area separating unit 60 obtains the type of each classified image, the position information where the area of the image appears, and the position information where the invalid area appears. The image is sent to the overall position / distortion correction unit 41 and the approximate difference extraction unit 42 of the difference extraction unit 40 and the image output unit 50. As described above, the present embodiment is characterized in that the images in the image area included in the effective area are classified by type.

  When the position information of the invalid area is sent, the overall position / distortion correction unit 41 in the difference extraction unit 40 determines the total position from the area (effective area) excluding the invalid area of both the pre-correction image and the post-correction image. Feature points for distortion correction are extracted (step 202). At this time, the overall position / distortion correction unit 41 extracts feature points from the image in the image area included in the effective area by a method suitable for the type of the image. For example, if the image is a character image, feature points are extracted in units of characters, and if the image is a line drawing image or a table ruled line image, an image portion of line intersection or bending is extracted as a feature point. Accordingly, an optimum feature point extraction method and processing parameters are adopted. When more accurate feature points are extracted in this way, the overall position / distortion correction unit 41 corrects the positional distortion of the corrected image using the extracted feature points of both images (step 104).

  Since the subsequent processing in the difference extraction unit 40 (steps 105 to 107) and the processing in the image output unit 50 (steps 108 to 109) may be the same as those in the first embodiment, description thereof will be omitted.

  In the present embodiment, the case where the present invention is applied to the original purpose of the difference extraction apparatus, which is a process for confirming a correction location and its contents by extracting the difference between the original document before correction and the original document after correction. explained. For this reason, in the present embodiment, the original before and after correction is read using the scanner 10, and the second image to be extracted as a difference between the pre-correction image corresponding to the first image and the first image. As a result, the modified image is generated. However, the difference extraction target is not necessarily limited to the originals before and after the correction, and may be simply applied to the originals to be compared. Further, instead of reading the document by the scanner 10 and generating image data, the image reading unit 20 may read each image data to be compared from the HDD 4 or obtain it via the network controller 12. Good.

  Further, in the present embodiment, it is assumed that the difference extraction device is formed by one computer, but may be realized by combining a plurality of computers and other devices. For example, in the present embodiment, the printer 9 and the scanner 10 are connected to the computer and the image reading unit 20 is configured. However, a printer and a scanner mounted on the image forming apparatus may be used.

It is the block block diagram which showed Embodiment 1 of the difference extraction apparatus which concerns on this invention. 2 is a hardware configuration diagram of a computer forming the difference extraction device in Embodiment 1. FIG. It is a block block diagram of the whole position and distortion correction | amendment part shown in FIG. 3 is a flowchart showing difference extraction processing in the first embodiment. FIG. 6 is a block configuration diagram showing a difference extraction device in a second embodiment. 10 is a flowchart illustrating difference extraction processing in the second embodiment.

Explanation of symbols

  1 CPU, 2 ROM, 3 RAM, 4 hard disk drive (HDD), 5 HDD controller, 6 mouse, 7 keyboard, 8 display, 9 printer, 10 scanner, 11 I / O controller, 12 network controller, 13 internal bus, 20 images Reading unit, 30 Invalid region specifying unit, 40 Difference extracting unit, 41 Overall position / distortion correcting unit, 42 Roughly difference extracting unit, 43 Divided region calculating unit, 44 Region dividing unit, 45 Local position / distortion correcting unit, 46 Detailed difference Extraction unit, 47 Difference composition unit, 50 Image output unit, 51 Invalid region incorporation unit, 60 Region separation unit, 411 First feature point extraction unit, 412 Second feature point extraction unit, 413 Corresponding feature point determination unit, 414 correction Coefficient calculation unit, 415 Position / distortion correction unit.

Claims (3)

Accepting means for accepting a first image and a second image that is a difference extraction target of the first image;
An invalid area specifying means for specifying, as an invalid area, an area including an image from which an effective processing result is not obtained when executing the difference extraction process from at least one of the first and second images;
Difference extraction processing means for performing difference extraction processing by comparing image areas obtained by removing invalid areas from the first and second images;
Image output means for outputting a difference extraction image obtained as a result of the processing by the difference extraction processing means in a manner capable of discriminating the invalid area specified by the invalid area specifying means ;
I have a,
When printing the difference extraction image, the image output means cuts out an image of the invalid area portion specified by the invalid area specifying means from each of the first and second images, and outputs each image of the cut out invalid area portion. Is printed separately from the difference extraction image . The difference extraction device according to claim 1,
The difference extraction processing means includes:
A feature point extraction unit that extracts a feature point from the image area excluding the invalid area of each of the first and second images;
A correction unit that corrects the position and distortion of at least one of the first and second images by associating the feature points extracted from the image area excluding the invalid area of each of the first and second images;
A difference extraction unit that performs difference extraction by comparing the first and second images after correction by the correction unit;
Have
The invalid area specifying means further specifies the type of the partial image included in the image and the area of the partial image from at least one of the first or second images,
The feature point extraction unit is characterized from a region of the partial image by a method according to the type of the specified partial image in an image region excluding an invalid region of at least one of the first or second images. A difference extraction apparatus characterized by extracting points. The computer,
Accepting means for accepting a first image and a second image to be extracted from the first image,
An invalid area specifying means for specifying, as an invalid area, an area including an image from which an effective processing result cannot be obtained when executing the difference extraction process from at least one of the first and second images;
Difference extraction processing means for performing difference extraction processing by comparing image regions obtained by removing invalid regions from the first and second images,
Image output means for outputting a difference extraction image obtained as a result of the processing by the difference extraction processing means in a manner capable of determining the invalid area specified by the invalid area specifying means ;
Function as
When printing the difference extraction image, the image output means cuts out an image of the invalid area portion specified by the invalid area specifying means from each of the first and second images, and outputs each image of the cut out invalid area portion. Is printed separately from the difference extraction image .

Images