JP2013156946A - Comic image data detecting device, comic image data detecting program, and comic image date detecting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a comic image data detecting device that can reduce a capacity of template data used for matching and can shorten a processing time of the matching when detecting comic image data.SOLUTION: A frame division information extracting unit 11 converts obtained comic image data into edge image data, and extracts frame division image data in which frame division lines remain and cartoon parts are removed from the edge image data. A plurality of frame division image data with different sizes are generated with the extracted frame division image data as reference, and the generated frame division image data are stored in a storage device 70 as frame division information. A matching unit 15 matches image data obtained from a Web site searched by a Web site searching unit 13 with the frame division image data.

Description

  The present invention relates to a technique for detecting image data matching or similar to comic image data, which is data obtained by converting a comic page to be detected into image data from image data existing on a Web site.

Conventionally, there has been disclosed a technique for detecting information with high illegality from search result information searched based on search conditions for information published on a network (see Patent Document 1).
The illegal information detection apparatus described in Patent Document 1 extracts feature amounts from search result information, and stores the extracted feature amounts in association with search result information. Furthermore, the similarity between each search result information is calculated using collation techniques, such as a text, an audio | voice, an image, and a moving image, and it is determined whether it is similar based on the calculated similarity. When the number of search result information determined to be similar exceeds a predetermined value, the detection result information is detected as highly illegal information.

  Further, for example, Patent Document 2 discloses a pattern matching device that performs matching between a template image of bitmap data and an image to be collated. In such a pattern matching apparatus, edge information is extracted from the bitmap data and the image to be collated. Next, a smoothing process is performed on both images that have become edge images to compensate for changes in the shape of both. In this way, matching processing is performed on the edge images in which the shape deformation portions are corrected.

JP 2009-199513 A JP 2008-159061 A

  For example, when the matching technique of Patent Document 2 is applied to the matching of the image of Patent Document 1, bitmap data (multi-tone image data) needs to be prepared as a template image. If there are few types of collation targets, it is sufficient to store and hold bitmap data as many as the number of collation targets. However, if there are many collation targets such as comic image data existing on a website, prepare as a template image. There are various types of bitmap data to be processed, and the data capacity becomes enormous. For this reason, it is necessary to delete unnecessary information from the bitmap data, such as making an edge image, to reduce the capacity of each template data. However, in order to ensure the target collation accuracy, it is necessary to extract an appropriate feature amount according to the collation target, and the time required for the matching process varies depending on the content of the extracted feature amount. However, until now, no detection method using appropriate template data has been proposed for detecting comic image data.

  The present invention has been made to solve such a problem, and its purpose is to reduce the capacity of template data and reduce matching time when detecting comic image data. To provide a detection device, a comic image data detection program, and a comic image data detection method.

  [Mode 1] In order to achieve the above object, the comic image data detection device according to mode 1 is data obtained by converting a preset page in a comic to be detected into image data from image data existing on a Web site. A comic image data detection apparatus for detecting image data that is identical or similar to comic image data, wherein the comic image data of the comic to be detected is at least the shape of each frame divided in the comic image data and each Frame allocation information extracting means for extracting frame allocation information, which is information related to frame allocation including information on the position of the frame, frame allocation information storage means for storing the frame allocation information extracted by the frame allocation information extracting means, Image data acquisition means for acquiring image data from a website, and image data acquired by the image data acquisition means Based on frame allocation information stored in the frame allocation information storage unit, matching means for matching the acquired image data with the frame allocation information, and the acquired image data based on a matching result of the matching means And detecting means for detecting image data that matches or is similar to the comic image data from the acquired image data based on the similarity between the image data and the comic image data.

With such a configuration, the frame allocation information extracting unit detects, from the comic image data of the comic to be detected, a frame including information on the shape of each frame and the position of each frame in the comic image data. When the division information is extracted, the extracted piece division information is stored by the piece division information storage means. On the other hand, when image data is acquired from the website by the image data acquisition unit, the acquired image is acquired based on the image data acquired by the matching unit and the frame division information stored in the frame division information storage unit. Matching between data and frame allocation information is performed. When matching is performed by the matching unit, the detection unit detects image data that matches or is similar to the comic image data from the acquired image data based on the similarity between the acquired image data and the comic image data based on the matching result. Is done.
This makes it possible to extract frame allocation information that is unique to comics and perform matching using the extracted frame allocation information, so matching can be performed while maintaining detection accuracy relatively high. The amount of template data to be used can be reduced.

  [Mode 2] Further, in the comic image data detection device according to mode 2, the frame division information extraction unit is configured to convert the comic image data into edge component included in the image of the comic image data. Frame image that is converted into edge image data that is image data indicating a distribution, and image data in which only edge components corresponding to the outer edge portion of each frame that has been divided are drawn from the edge image data as the frame division information. The split image data is extracted.

  With such a configuration, image data obtained by removing excess information from comic image data and including information on the shape, position, and size of each frame can be extracted as frame allocation information. Therefore, the data amount of the frame allocation information can be greatly reduced. In addition, since the frame division information is information unique to comics, it is information with relatively high detection accuracy as template data at the time of matching.

[Mode 3] Further, in the comic image data detection device according to mode 3, the frame division information extraction unit is configured to perform the frame division on the extracted frame division image data corresponding to the comic image data. As the split information, a plurality of frame split image data having different image sizes configured by changing the image size while maintaining the aspect ratio is generated.
With such a configuration, it is possible to generate a plurality of frame division image data having different sizes as frame division information while maintaining the aspect ratio.
This makes it possible to perform matching using frame-divided image data of an appropriate size corresponding to the size of image data acquired from a website, the number of comic pages included in the image data, and the like. . Therefore, detection accuracy can be improved.

[Mode 4] Further, in the comic image data detection device according to mode 4, the detection means is detected as the matching or similar image data for each Web site in the configuration of any one of modes 1 to 3. Based on the image data, it is determined whether or not the detected image data meets a preset narrowing condition. If it is determined that the image data matches, there is a possibility that the image data has been illegally disclosed. It will be detected as comic image data,
It comprises site information storage means for storing comic image data detected by the detection means, which may be illegally disclosed, and information on a website that discloses the image data in association with each other. And

  In such a configuration, depending on the setting of the filtering conditions, there is a possibility that the site may have illegally released comic image data (hereinafter referred to as illegal candidate site), or that it may have been illegally released. Certain comic image data (hereinafter referred to as illegal candidate image data) can be detected appropriately. In addition, for example, it is possible to set a narrowing condition so as to exclude a legitimate site so as not to detect a website that regularly conducts electronic mail order of electronic comics as an illegal candidate site.

[Mode 5] Further, in the comic image data detection apparatus according to mode 7, the detection means is configured so that the detection means has image data detected as the matching or similar image data as a preset number of pages. When it is determined that the above comic image data is included, it is determined that the narrowing condition is satisfied.
With such a configuration, for example, a site where, for example, an image of pages 1 to 3 is published as a comic introduction page, such as a regular electronic comic sales site, is excluded from detection targets. It is possible to prevent a large number of sites to be detected (listed).

[Mode 6] On the other hand, in order to achieve the above object, the comic image data detection program according to mode 6 is data obtained by converting a preset page in a comic to be detected into image data from image data existing on a website. A comic image data detection program for detecting image data that is the same as or similar to comic image data, the shape of each frame divided at least in the comic image data from the comic image data of the comic to be detected And frame allocation information extracting means for extracting frame allocation information, which is information relating to frame allocation including information on the position of each frame, image data acquisition means for acquiring image data from the Web site, and image data acquisition means Frame allocation storing image data and frame allocation information extracted by the frame allocation information extraction means Based on the frame division information stored in the information storage means, matching means for matching the acquired image data with the frame division information, and the acquired image data based on the matching result of the matching means and the A program for causing a computer to execute a function realized as detection means for detecting image data that matches or is similar to the comic image data from the acquired image data based on the similarity to the comic image data. And
With such a configuration, when the program is read by the computer and the computer executes the process according to the read program, the same operation and effect as those of the comic image data detection device according to mode 1 can be obtained.

[Mode 7] In order to achieve the above object, the comic image data detection method of mode 7 is a data obtained by converting a preset page in a comic to be detected into image data from image data existing on a website. A comic image data detection method for detecting image data that matches or is similar to comic image data, wherein the shape of each frame divided at least in the comic image data from the comic image data of the comic to be detected And a frame allocation information extracting step for extracting frame allocation information, which is information related to the frame allocation including the position information of each frame, and a frame allocation information storage means for storing the frame allocation information extracted in the frame allocation information extraction step; , An image data acquisition step of acquiring image data from the Web site, and the image data acquisition step Based on the image data acquired in step 1 and the frame allocation information stored in the frame allocation information storage step, a matching step for matching the acquired image data with the frame allocation information, and based on a matching result of the matching step And a detection step of detecting image data that matches or is similar to the comic image data from the acquired image data based on the similarity between the acquired image data and the comic image data.
Thereby, the same operation and effect as those of the comic image data detection apparatus according to the first aspect can be obtained.

  As described above, according to the present invention, information related to frame division in comic image data is extracted and stored as data (template data) used for matching, so detection accuracy is relatively high. The effect that the memory capacity required for storing the template data can be reduced while maintaining the accuracy is obtained. In addition, before manually detecting websites that illegally publish comic image data, it is possible to narrow down the websites that may illegally publish comic image data. The effect of reducing labor can be obtained.

It is a figure which shows the outline | summary of a comic image data detection system. 3 is a block diagram showing a hardware configuration of comic image data detection apparatus 100. FIG. 3 is a block diagram illustrating an example of a functional configuration of a comic image data detection apparatus 100. FIG. (A) is a figure which shows an example of comic image data, (b) is a figure which shows an example of the frame division | segmentation image data extracted from the comic image data of (a). It is a flowchart which shows an example of the process sequence of a frame allocation information extraction process. It is a flowchart which shows an example of the process sequence of a comic image data detection process. (A)-(e) is a figure which shows an example of the frame division content of frame division image data. It is a figure which shows an example of frame size image data of multiple sizes.

Hereinafter, embodiments of a comic image data detection device, a comic image data detection program, and a comic image data detection method according to the present invention will be described with reference to the drawings. FIGS. 1-8 is a figure which shows embodiment of the comic image data detection apparatus, comic image data detection program, and comic image data detection method concerning this invention.
(Constitution)
FIG. 1 is a diagram showing an outline of a comic image data detection system according to an embodiment of the present invention.
As shown in FIG. 1, the comic image data detection system 1 includes a comic image data detection device 100, Web servers 200_1 to 200_N (N is a natural number of 4 or more), and client terminals 300_1 to 300_M (M is 2 or more). Natural number).

The comic image data detection apparatus 100, the Web servers 200_1 to 200_N, and the client terminals 300_1 to 300_M are connected to each other via a known Internet 400 so that data communication can be performed.
The comic image data detection apparatus 100 has a function of detecting comic image data, which is data obtained by converting comic pages into image data, from image data published on a website provided by the web servers 200_1 to 200_N. . Furthermore, the comic image data detection apparatus 100 has a function of detecting, as an illegal candidate site, a Web site that may illegally disclose comic image data of a comic to be detected.

The Web servers 200_1 to 200_N are configured from servers under the management of an Internet provider or the like, and are servers that provide a large number of Web sites that publish content such as text, audio, images, and moving images. Hereinafter, when it is not necessary to distinguish, the Web servers 200_1 to 200_N are simply referred to as the Web server 200.
The client terminals 300_1 to 300_M are terminals that can access a website provided by the web server 200 via the Internet 400, and browse and download various contents published on the accessed website. is there. Hereinafter, when there is no need to distinguish, the client terminals 300_1 to 300_M are simply referred to as client terminals 300.

Next, the hardware configuration of the comic image data detection apparatus 100 will be described with reference to FIG. FIG. 2 is a block diagram showing a hardware configuration of comic image data detection apparatus 100.
As shown in FIG. 2, the comic image data detection apparatus 100 includes a CPU (Central Processing Unit) 60, a RAM (Random Access Memory) 62, a ROM (Read Only Memory) 64, and an input / output interface (I / F). 66 and a computer system including a bus 68 is provided. The CPU 60, RAM 62, ROM 64 and I / F 66 are connected to a bus 68, and data can be transmitted and received between these connected devices via the bus 68.

The comic image data detection device 100 is further connected to a storage device 70, a display device 72, an input device 74, an image scanner 76, and a NIC (Network Interface Card) 78 via an I / F 66. .
The CPU 60 reads various dedicated computer programs stored in advance in the ROM 64 or the storage device 70 into the RAM 62, and makes use of various resources in accordance with instructions described in the programs read into the RAM 62 to perform frame allocation information extraction processing to be described later. And comic image data detection processing.

The storage device 70 stores data necessary for the frame division information extraction process and the comic image data detection process, such as various dedicated computer programs and data used in the execution of the program.
Specifically, the storage device 70 stores the frame division information extracted in the frame division information extraction process as template data used for the matching process.

Furthermore, the storage device 70 detects comic image data (illegal candidate image data) that may be illegally detected in the comic image data detection process, and a Web site that discloses the illegal candidate image data ( Information on illegal candidate sites) is stored.
Here, the storage device 70 is a magnetic storage type storage device such as an FD drive or HD drive, an optical read / write type storage device such as a CD drive, a DVD drive or a BRD drive, or a magnetic storage type / optical such as an MO drive. Any storage device can be used as long as it can read stored data and write data on a computer, such as a conventional reading method storage device.

The display device 72 is a well-known display device such as a CRT display or a liquid crystal display, and publishes the illegal candidate image data detected in the comic image data detection process and the illegal candidate image data in response to a command from the CPU 60. Display images such as information on illegal candidate sites.
The input device 74 is a human interface device such as a keyboard, a mouse, an operation button, or a touch-type input device (such as a touch panel), and receives an input corresponding to a user operation.

In response to a command from the CPU 60, the image scanner 76 optically reads a document (comic page in the present embodiment) set in the document reading unit, and reads image data of the read document (comic image in the present embodiment). Data). Then, the generated comic image data is sequentially transmitted to the comic image data detection apparatus 100 via the I / F 66.
The NIC 78 is a device for communicating with an external device on a network such as a LAN, WAN, or the Internet, and is a known network adapter generally called a LAN card or the like. The NIC 78 is connected to a router, a hub, or the like via a network cable L, and connects the comic image data detection device 100 to a network (the Internet in this embodiment) via these. It may be configured to connect to a network via a wireless LAN access point, such as a known wireless LAN.

Thereby, in this embodiment, it is possible to acquire image data published on each Web site via the NIC 78 from the Web site provided by the Web server 200 connected to the network.
The comic image data detection apparatus 100 is configured to be mounted on, for example, a well-known personal computer.

Next, based on FIG.3 and FIG.4, the function structure of the comic image data detection apparatus 100 is demonstrated. FIG. 3 is a block diagram illustrating an example of a functional configuration of the comic image data detection apparatus 100. FIG. 4A is a diagram illustrating an example of comic image data, and FIG. 4B is a diagram illustrating an example of frame divided image data extracted from the comic image data in FIG.
As illustrated in FIG. 3, the comic image data detection device 100 includes a comic image data acquisition unit 10, a frame division information extraction unit 11, a data communication unit 12, a website search unit 13, and an image as function configuration units. The data acquisition part 14, the matching part 15, the determination part 16, and the detection part 17 are provided.
The comic image data acquisition unit 10 transmits a document reading instruction to the image scanner 76 in response to a data acquisition instruction from the operator via the input device 74.

  In the present embodiment, the operator sets a comic page to be detected in the document reading unit of the image scanner 76. Accordingly, the image scanner 76 optically reads the set comic page by the image sensor in response to the reading instruction from the comic image data acquisition unit 10 and generates image data (comic image data) of the read page. To do. Then, the image scanner 76 transmits the generated comic image data to the comic image data detection device 100 via the I / F 66. In addition, the operator inputs information on the comic (title, author name, volume, etc.) set in the document reading unit to the comic image data detection device 100 via the input device 74.

  The image scanner 76 includes a paper feed tray on which a plurality of originals can be placed, and has a function of automatically setting the originals placed on the paper feed tray one by one in the original reading unit each time the originals are read. Suppose you have it. In this case, the comic image data acquisition unit 10 transmits a reading instruction for continuous reading to the image scanner 76, whereby a plurality of originals placed on the paper feed tray are successively fed to the image scanner 76 one by one. Can be read.

  The comic image data acquisition unit 10 acquires comic image data transmitted from the image scanner 76. Then, the acquired comic image data is stored in a preset storage area of the RAM 62. At this time, the acquired comic image data is stored in association with the input comic information. Further, when the comic image data acquisition unit 10 completes the acquisition of the comic image data, the comic image data acquisition unit 10 notifies the frame allocation information extraction unit 11 that the acquisition of the comic image data has been completed.

The frame allocation information extraction unit 11 reads comic image data stored in the RAM 62 in response to the notification from the comic image data acquisition unit 10, and extracts frame allocation information from the read comic image data.
For example, it is assumed that comic image data CI1 having the drawing content shown in FIG.
As shown in FIG. 4A, the image of the comic image data CI1 is composed of four rectangular frames F1 to F4 and a margin portion MA that is a region portion outside the frames F1 to F4. The frame F1 is composed of a frame dividing line E1 which is a line segment forming the outer edge portion thereof, and a comic P1 drawn in an area surrounded by the frame dividing line E1, and the frame F2 is a frame forming the outer edge portion thereof. It is composed of a dividing line E2 and a comic P2 drawn in an area surrounded by the frame dividing line E2. Further, the frame F3 includes a frame dividing line E3 that forms an outer edge portion thereof, and a comic P3 drawn in an area surrounded by the frame dividing line E3. The frame F4 forms a frame dividing line E4 that forms the outer edge portion thereof. And the comic P4 drawn in the area surrounded by the frame dividing line E4.

The frame allocation information extraction unit 11 first converts the read comic image data CI1 into edge image data using a known edge extraction method.
Here, as an edge extraction method, for example, a well-known differential operator such as a Prewitt operator, a Sobel operator, a Laplacian, or the like is used to differentiate (edge emphasis) comic image data, and after edge emphasis, binarization set in advance is performed. There is a method of binarization using a threshold value. In addition, as a binarization method, there are a method of using one fixed value as a binarization threshold, a method of dividing an image of comic image data into a plurality of blocks, and using a different binarization threshold for each block, and the like. . The binarization threshold value may be a value input by the operator, or may be obtained by calculation each time from a binarized image histogram or the like.

If the comic page is a color page, the comic image data may be converted to grayscale data and then edge extraction or the like may be performed.
Further, the binarized image data may be edge image data, but after binarization, a linearization process for linearizing the edge portion may be performed. Moreover, you may perform a noise removal process (for example, expansion | swelling, shrinkage | contraction etc.) as needed.
As the linearization processing, for example, there is a method of linearizing an edge portion using a known Bresenham line segment drawing algorithm.

The edge image data generated in this way is composed of, for example, binary values where the pixel value of the edge portion is “1” and the pixel value of the portion other than the edge is “0”. In the example of comic image data shown in FIG. 4A, the content of the edge image data is substantially the same because the comic content is centered on a line drawing that is not screen-tone or solid.
Next, the frame division information extraction unit 11 leaves the frame division line, which is a line segment forming each frame, from the edge image data, and extracts the image data of the content from which the cartoon portion is removed.

  For example, in the example of FIG. 4A, the frame dividing lines E1 to E4 that form the respective frames F1 to F4 are left, and the image data of the content from which the comic portions P1 to P4 are removed is extracted. As a result, image data FI1 (hereinafter referred to as frame division image data FI1) having the image content shown in FIG. 4B is extracted (generated). As shown in FIG. 4B, the frame division image data FI1 has image contents in which only the frame division lines E1 to E4 of the frames F1 to F4 are drawn on the page. As shown in FIG. 4B, the frame division image data is data having information on the shape, size, and position of each frame on the page paper. In the example of FIG. 4A, each frame forms a closed region by a frame dividing line. However, the frame dividing line, for example, a single horizontal line that bisects the paper surface is the outer edge portion of the paper surface. In some cases, each frame is configured as an edge of the area. In this case, the blank portion is eliminated. Further, the frame dividing line is not limited to a line orthogonal to the vertical and horizontal directions of the page, but may be a non-orthogonal line such as an oblique line. Further, the shape of the closed region constituted by the frame dividing lines is not limited to a rectangle, but may be another shape such as a trapezoid or a triangle.

Furthermore, in this embodiment, the frame division information extraction unit 11 uses the extracted frame division image data as a reference (100 [%]) and maintains the aspect ratio, for example, 10 [%] to 200 [%]. In this range, image data obtained by reducing and enlarging the reference frame-cut image data in increments of 10 [%] is generated. Then, the frame divided image data in 10 [%] increments from 10 [%] to 200 [%] is compressed by a compression method having reversibility such as run length compression, and these compressed data correspond to comic information. Then, it is stored in the storage device 70. In the present embodiment, this multi-size frame division image data is frame division information. Since the frame division image data includes only a single color and data of the frame division line portion of each frame as the drawing portion data, the data amount can be greatly reduced. Note that the range of reduction and enlargement is not limited to the range of 10 [%] to 200 [%], but may be other ranges, and is not limited to 10 [%] increments, and may be other increments. .
The data communication unit 12 has a function of controlling transmission of various data to an external device and reception of various data from the external device via the NIC 78.

  Specifically, the data communication unit 12 has a function of transmitting data to an external device such as the Web server 200 connected to the Internet 400 via the NIC 78 in response to a data transmission request from each functional configuration unit. ing. Further, the data communication unit 12 has a function of outputting data received from an external device connected to the Internet 400 to each corresponding functional component via the NIC 78.

In response to the patrol start instruction input from the operator via the input device 74, the website search unit 13 receives the comic title included in the comic information associated with the frame-cut image data from the storage device 70, Read information such as author name and volume. Then, the web site corresponding to the search word is searched from the web site provided by the web server 200 connected to the Internet 400, using the read word indicating the author, author name, volume number, etc. as a search word.
Specifically, the website search unit 13 accesses, for example, a preset website that provides a reliable search service via the data communication unit 12. Then, a search word is transmitted to this Web site via the data communication unit 12, and the Web site is searched using a search service.

Then, the website search unit 13 stores the information (URL) of the searched website in the storage device 70 in association with the search word. Thereafter, the website search unit 13 notifies the image data acquisition unit 14 that the search for the website has been completed.
Note that the information is not limited to a configuration that stores information on all searched websites, and only information on a predetermined upper limit number of websites is stored in consideration of the case where the number of searches (hits) is enormous. It is good also as a structure.

Further, in the present embodiment, the website search unit 13 lists, in advance, websites that have permission from a copyright holder, such as websites that regularly sell electronic comics online. Is generated and is excluded from the search result when a Web site in the list is searched.
In response to the notification from the website search unit 13, the image data acquisition unit 14 responds to each website via the data communication unit 12 based on the website information (URL) stored in the storage device 70. Access in order. Then, image data published on the accessed website is acquired via the data communication unit 12, and the acquired image data is stored in the storage device 70.

  Further, when the image data acquisition unit 14 acquires the image data from all the searched Web sites, the image data acquisition unit 14 notifies the matching unit 15 that the acquisition of the image data is completed. Further, the image data acquisition unit 14 temporarily stops the acquisition of the image data every time the preset upper limit number of image data is acquired in response to the case where the number of image data to be acquired is large. 15, a matching instruction is transmitted. On the other hand, the notification of the completion of matching is received from the matching unit 15, and in response to receiving this notification, the acquisition of the stopped image data is resumed.

  Note that the present invention is not limited to the above-described configuration in which the website search unit 13 searches for a website using a search word using a search service, and acquires image data from the website searched by the image data acquisition unit 14. Other configurations may be used. For example, an acquisition condition is set in advance without performing a search for a website using a search word, and the image data acquisition unit 14 changes the acquired acquisition condition from the website provided by the Web server 200 that matches the set acquisition condition. A configuration may be adopted in which all matching image data is acquired. As acquisition conditions, for example, an area where the Web server 200 is located (Japan, the United States, etc.), an update date of image data, and the like are set. Also, as the update date and time, for example, the date and time after the sale date of the comic corresponding to the frame allocation information is set.

In response to an acquisition completion notification or a matching instruction from the image data acquisition unit 14, the matching unit 15 stores image data (hereinafter referred to as acquired image data) stored in the storage device 70, It has a function for matching with split image data.
Specifically, the matching unit 15 converts the acquired image data into edge image data as in the frame division information extraction process, and leaves only the image portion estimated as the frame dividing line from the converted edge image data. (Hereinafter referred to as “acquired frame divided image data”) is extracted. The extracted frame division image data is subjected to template matching using each frame division image data stored in the storage device 70 as a template image.
In addition, it is good not only as a structure which matches after converting the acquired image data into edge image data, For example, it is good also as other structures, such as a structure which matches using acquired image data as it is.

  Acquired image data can include one image data including one page of comics, or two pages or more. In the present embodiment, matching is performed for the case of including one page and the case of including two spread pages or two pages in a disjoint page order. At that time, according to the size of the acquired image data, matching is performed by selecting two types of image data having the optimum size from the divided image data of each size from 10 [%] to 200 [%]. For example, for two pages of spread or disjoint pages, the half-sized frame division image data is selected with respect to the size when one page is included.

In addition, as a matching method, for example, the resolution of the frame-cut image data and the acquired frame-cut image data is made the same, the difference between the pixel values of each pixel at the same pixel position is calculated, and the absolute value of this difference value is calculated. Calculate the sum. Since the frame-cut image data and the acquired frame-cut image data are image data binarized (in this case, binarized to “0” and “1”), they are the same when the difference is calculated. The calculation result is “0” in the pixel value portion, and “1” or “−1” in the different pixel value portion.
The matching unit 15 outputs the sum of the difference values between the frame-cut image data and the acquired frame-cut image data, which is a matching result, to the determination unit 16.

The determination unit 16 compares the total acquired from the matching unit 15 with a preset matching threshold, and if the total is equal to or less than the matching threshold, the frame-cut image data and the acquired frame-cut image data match or are similar. If the sum is larger than the matching threshold, it is determined that the frame-cut image data and the acquired frame-cut image data are different (not similar) image contents. In the case of this calculation method, the smaller the sum is, the more similar the two are. When the sum is “0”, the two are the same. That is, this sum indicates the degree of similarity between the frame-cut image data and the acquired frame-cut image data.
The determination unit 16 generates determination result information in which the determination result is associated with information on a website that publishes the corresponding acquired image data, and outputs the generated determination result information to the detection unit 17.

Based on the determination result information from the determination unit 16, the detection unit 17 counts the number of image data included in the acquired frame-allocated image data determined to have the same or similar image content for each Web site. When all the determination result information is acquired for each detected Web site, the count value is compared with a preset count threshold value. As a result of this comparison, if the count value is equal to or greater than the count threshold, the acquired image data corresponding to the counted acquired frame-cut image data is comic image data (illegal image data) that may be illegally disclosed. Is determined. On the other hand, when the count value is less than the count threshold, it is determined that the acquired image data corresponding to the acquired acquired frame-sliced image data is not comic image data that may have been illegally disclosed.
Here, the count value indicates the number of image data (pages) determined to be identical or similar to the comic image data of the comic to be detected used in the matching process.

  The count threshold value is a threshold value for excluding comic image data published as an introduction page by an authorized sales site from illegal image data, for example. In other words, for example, when about 1 to 3 pages of comics are excerpted and introduced, the threshold is set to the upper limit number of pages so as not to make an illegal decision. From another point of view, it can be said that the number of illegal candidate sites is a threshold for narrowing down the number so as not to hinder the crackdown.

  When the detection unit 17 determines that there is a possibility of being illegally disclosed, the detection unit 17 discloses at least a part of the comic image data determined to be illegally disclosed and the illegal candidate image data. The detection result information for illegal use, which is information that associates the information of the Web site (illegal candidate site) with the comic information, is generated. The generated illegal detection result information is stored in the storage device 70. On the other hand, if it is determined that there is no possibility that the information has been illegally disclosed, the detection unit 17 uses the information for the website, the information indicating that the information is not illegal, and the comic information for non-illegal use. The detection result information is generated. Then, the generated illegal detection result information is stored in the storage device 70. Further, the detection unit 17 deletes the acquired image data corresponding to the website determined to be not illegal from the storage device 70. Note that when it is determined that it is not illegal, the information on the website and the acquired image data may be deleted.

Further, the detection unit 17 notifies the detection result display unit 18 that the detection result information has been stored (or updated).
The detection result display unit 18 reads the detection result information from the storage device 70 in response to the notification from the detection unit 17 and displays an image indicating the detection result on the display device 72 based on the read information.

(Frame division information extraction process)
Next, based on FIG. 5, the processing procedure of the frame division information extraction processing executed by the CPU 60 according to the dedicated computer program stored in the ROM 64 will be described. FIG. 5 is a flowchart illustrating an example of a processing procedure of frame allocation information extraction processing.
When the frame division information extraction process is executed in the CPU 60, first, the process proceeds to step S100 as shown in FIG.

In step S <b> 100, the comic image data acquisition unit 10 determines whether or not there has been a data acquisition instruction from the operator via the input device 74. As a result, when it is determined that there is a data acquisition instruction from the operator (Yes), the process proceeds to step S102. On the other hand, when it is determined that there is no data acquisition instruction from the operator (No), the determination process is repeated until there is a data acquisition instruction.
When the process proceeds to step S102, the comic image data acquisition unit 10 acquires the comic image data transmitted from the image scanner 76, and the process proceeds to step S104.

  Specifically, the comic image data acquisition unit 10 transmits a reading instruction to the image scanner 76 in response to the data acquisition instruction. In response to the reading instruction, the image scanner 76 optically reads the comic page (original), and generates comic image data of the read page. Then, the comic image data acquisition unit 10 acquires comic image data transmitted from the image scanner 76 via the I / F 66.

In step S104, the comic image data acquisition unit 10 stores the comic image data acquired in step S102 in the RAM 62 in association with comic information input by an operator separately, and the process proceeds to step S106.
If the comic image data acquisition unit 10 determines in step S106 that the comic image data has been stored, the comic image data acquisition unit 10 is notified of the acquisition of comic image data, and the process proceeds to step S108.

In step S108, the frame allocation information extraction unit 11 reads the comic image data stored in the RAM 62, and proceeds to step S110.
In step S110, the frame allocation information extraction unit 11 executes frame allocation information extraction processing for extracting frame allocation information from the comic image data read in step S108, and the process proceeds to step S112.

  In the frame division information extraction process, the frame division information extraction unit 11 first converts the read comic image data into edge image data. Next, the frame image data of the contents obtained by removing the cartoon portion drawn inside the outer edge portion of each frame is extracted while leaving the frame dividing line forming the outer edge portion of each frame divided from the edge image data. . Further, the frame division information extraction unit 11 uses the size of the extracted frame division image data as a reference (100 [%]), and sets the reference frame division in the range of 0 [%] to 200 [%] in increments of 10 [%]. A plurality of frame-divided image data having different sizes is generated by reducing and enlarging the image data while maintaining the aspect ratio. A set of the generated frame divided image data of each size and the extracted reference frame divided image data as a set and associated with the comic information is the frame divided information.

Further, the frame division information extraction unit 11 sequentially performs frame division information extraction processing on all comic image data stored in the RAM 62 and sequentially deletes comic image data for which extraction processing has been completed from the RAM 62.
In step S112, the frame allocation information extraction unit 11 stores the frame allocation information generated in the frame allocation information extraction process in step S110 in association with the comic information in the storage device 70, and the process proceeds to step S100.
In this embodiment, the frame division information extraction unit 11 compresses a plurality of frame division image data by a compression method having reversibility such as run-length compression, and associates the compressed frame division image data with comic information. Remember.

(Comic image data detection processing)
Next, a processing procedure of comic image data detection processing executed by the CPU 60 according to a dedicated computer program stored in the ROM 64 will be described with reference to FIG. FIG. 6 is a flowchart illustrating an example of a processing procedure of comic image data detection processing.
When the comic image data detection process is executed in the CPU 60, the process first proceeds to step S200 as shown in FIG.

  In step S <b> 200, the website search unit 13 determines whether there is a patrol start instruction from the operator via the input device 74. As a result, when it is determined that there is a patrol start instruction (Yes), the process proceeds to step S202. On the other hand, when it is determined that there is no patrol start instruction (No), the determination process is repeated until there is a patrol start instruction.

In step S202, the Web site search unit 13 reads a search word from the storage device 70, and proceeds to step S204.
Specifically, the Web site search unit 13 reads out word information such as a title and an author name included in the comic information from the comic information stored in the storage device 70 in association with the comic image data as a search word.

In step S204, the website search unit 13 searches for a website corresponding to the search word from the website provided by the web server 200 connected to the Internet 400 using the search word read in step S202. The process proceeds to step S206.
In step S206, the website search unit 13 determines whether a website corresponding to the search word has been searched. As a result, when it is determined that the website corresponding to the search word has been searched (Yes), the process proceeds to step S208. On the other hand, if it is determined that the Web site corresponding to the search word has not been searched (No), the process proceeds to step S244.

When the process proceeds to step S208, the website search unit 13 stores the information (URL) of the searched website in the storage device 70, and the process proceeds to step S210.
When the process proceeds to step S210, the image data acquisition unit 14 sequentially accesses the website corresponding to the information stored in the storage device 70, acquires the image data published on the accessed website, The process proceeds to step S212.
In step S212, the image data acquisition unit 14 stores the image data acquired in step S210 in the storage device 70, and the process proceeds to step S214.

In step S214, the image data acquisition unit 14 determines whether the number of acquired image data has reached a preset upper limit number. As a result, when it is determined that the upper limit number has been reached (Yes), the matching unit 15 is notified that the upper limit number has been reached, the process proceeds to step S216, and it is determined that the upper limit number has not been reached ( No) moves to step S242.
When the process proceeds to step S216, the matching unit 15 reads the acquired image data and the frame-divided image data corresponding to the acquired image data from the storage device 70, and the process proceeds to step S218.
In step S218, the matching unit 15 executes a matching process between the acquired image data read in step S216 and the frame division image data, and the process proceeds to step S220.

  In the matching process, the matching unit 15 first converts the acquired image data into edge image data, leaves the outer edge portion of the image portion estimated from each frame divided from the edge image data, and other drawing portions. The acquired frame-sliced image data with the content removed is extracted. Next, frame division image data having a size suitable for the size of the acquired frame division image data is selected from a plurality of frame division image data having different sizes. Then, the absolute value of the difference value between the pixel value of each pixel of the selected frame-cut image data and the pixel value of each pixel at the same pixel position of the acquired frame-cut image data is sequentially calculated. Then, the sum of these absolute values is calculated, and the calculated sum is sequentially output to the determination unit 16 as a matching result.

  In step S220, the determination unit 16 compares the total acquired from the matching unit 15 with a preset matching threshold, and based on the comparison result, the acquired image data matches or is similar to the comic image data. It is determined whether it is data. Then, determination result information is generated based on the determination result, the generated determination result information is output to the detection unit 17, and the process proceeds to step S222. In the present embodiment, information on the number of image data (hereinafter referred to as similar image data) determined to be coincident or similar included in the acquired image data is also added to the determination result information.

In step S222, based on the determination result information acquired from the determination unit 16, the detection unit 17 determines whether the matched acquired image data includes similar image data. As a result, when it is determined that similar image data is included (Yes), the process proceeds to step S224, and when it is determined that similar image data is not included (No), the process proceeds to step S226.
When the process proceeds to step S224, the detection unit 17 counts the number of similar image data (pages) included in the acquired image data for the corresponding Web site, and the process proceeds to step S226.

In step S226, the matching unit 15 determines whether the matching process has been completed for all acquired image data. As a result, when it is determined that the matching process is completed (Yes), the process proceeds to step S228, and when it is determined that the matching process is not completed (No), the process proceeds to step S218.
When the process proceeds to step S228, the image data acquisition unit 14 determines whether there is unacquired image data. When it is determined that there is unacquired image data (Yes), the process proceeds to step S210. If it is determined that there is no unacquired image data (No), the process proceeds to step S230.

When the process proceeds to step S230, the detection unit 17 compares the count value of similar image data with a preset count threshold value for each Web site, and the process proceeds to step S232.
In step S232, the detection unit 17 determines whether or not there is a website that matches the narrowing-down conditions based on the comparison result in step S230. As a result, when it is determined that there is a website that matches the narrowing-down conditions (Yes), the process proceeds to step S234.

On the other hand, if it is determined that there is no Web site that matches the filtering conditions (No), the process proceeds to step S236.
In the present embodiment, when the count value of the similar image data is equal to or greater than the count threshold, it is determined that the narrowing condition is met, and when the count value of the similar image data is less than the count threshold, It is determined that it does not match.
When the process proceeds to step S234, the detection unit 17 includes a part of the acquired image data determined to meet the filtering condition, information on illegal candidate sites that disclose the image data, and comic information. The detection result information for illegal use is generated, and the process proceeds to step S238.

On the other hand, when the process proceeds to step S236, the detection unit 17 generates illegal detection result information including information indicating that the site is not an illegal candidate site, information on the website, and comic information. The process proceeds to step S238.
In step S238, the detection unit 17 stores the detection result information generated in step S234 or S236 in the storage device 70, and the process proceeds to step S239.

In step S239, the detection unit 17 determines whether there is a Web site for which detection result information has not been generated. As a result, when it is determined that there is a Web site for which detection result information has not been generated (Yes), the process proceeds to step S230. On the other hand, if it is determined that there is no Web site for which detection result information has not been generated (No), the detection result display unit 18 is notified that the processing of storing (or updating) detection result information has been completed, and step S240 is performed. Migrate to
In step S240, the detection result display unit 18 reads the detection result information from the storage device 70, displays the detection result on the display device 72 based on the read detection result information, and proceeds to step S240.

  On the other hand, when it is determined in step S214 that the preset upper limit number has not been reached and the process proceeds to step S238, the image data acquisition unit 14 determines whether or not acquisition of image data is completed. As a result, when it is determined that the acquisition of the image data is completed (Yes), the completion is notified to the matching unit 15, and the process proceeds to step S216. On the other hand, if it is determined that the image data acquisition has not been completed (No), the process proceeds to step S210.

  In step S206, if the website corresponding to the search word is not searched and the process proceeds to step S240, the website search unit 13 determines whether there is a search word for another comic. As a result, when it is determined that there is a search word for another comic (Yes), the process proceeds to step S204, and when it is determined that there is no search word for another comic (No), the process proceeds to step S200.

(Operation)
Next, based on FIGS. 7-8, operation | movement of the comic image data detection apparatus 100 of this embodiment is demonstrated.
Here, FIGS. 7A to 7E are diagrams illustrating an example of the frame division contents of the frame division image data. Further, FIG. 8 is a diagram illustrating an example of a plurality of frames of divided image data.
First, the operation of the frame division information extraction process in the comic image data detection apparatus 100 will be described.

  Now, in the comic image data detection apparatus 100, it is assumed that the operator operates the input device 74 and inputs a data acquisition instruction. Thereby, the comic image data acquisition unit 10 outputs a reading instruction to the image scanner 76 via the I / F 66 in response to the input of the data acquisition instruction (Yes in step S100). At that time, the operator sets a page selected and separated in advance from the comic to be detected in the document reading section or the paper feed tray of the image scanner 76. This page is selected based on the matching accuracy at the time of matching processing, such as pages that are characteristic frame divisions in comics to be detected and pages that are closed areas surrounded by frame dividing lines. It is desirable to consider the extraction accuracy at the time of split information extraction. In addition, the operator inputs comic information such as the title, author name, and volume of the comic to be detected in advance.

The image scanner 76 optically reads a comic page set in the document reading unit, generates comic image data, and transmits the generated comic image data to the comic image data detection device 100 via the I / F 66. .
Then, the comic image data acquisition unit 10 acquires comic image data transmitted from the image scanner 76 (step S102), and stores the acquired comic image data in the RAM 62 in association with comic information (step S104). Thereafter, the comic image data acquisition unit 10 transmits a comic image data acquisition notification to the frame division information extraction unit 11 (step S106).

  When receiving the acquisition notification from the comic image data acquisition unit 10, the frame allocation information extraction unit 11 reads out comic image data from the RAM 62 (step S108), and extracts frame allocation information from the read comic image data. An extraction process is executed (step S110).

First, the frame allocation information extraction unit 11 performs edge enhancement processing by, for example, a known Sobel operator on the read comic image data. Next, the edge-enhanced data is binarized by a preset binarization threshold value to generate binarized edge image data. Further, the frame division information extraction unit 11 performs a process of linearizing the edge portion on the edge image data using, for example, a known Bresenham line segment drawing algorithm. After that, from the edge image data in which the edge portion is linearized, the frame dividing image data having the content in which the frame dividing line portion is left and the cartoon portion is removed is extracted according to a preset extraction rule. For example, as an extraction rule, a rule is set such that, for a portion where a closed region is formed by line segments in edge image data, an outer edge portion of the closed region is left and a drawing portion in the closed region is removed.
In this manner, it is assumed that the frame division image data illustrated in FIGS. 7A to 7E is extracted as the frame division information.

  Here, FIG. 7A is the same as FIG. 4B, and FIG. 7B is an example in which each frame has a rectangular shape similar to FIG. 7A. It is. FIG. 7C illustrates an example in which the shape of each frame is a trapezoid, and FIG. 7D illustrates an example in which a trapezoid and a rectangle are mixed as the shape of each frame. FIG. 7E shows an example in which frame division is performed with the end portion of the page as a part of the outer edge portion. In the example of FIG. 7E, it is difficult to determine each frame because the page edge is a frame dividing line.

The frame layout information extracting unit 11 further sets each image in increments of 10 [%] in the range of 10 [%] to 200 [%] with the image size of each extracted frame layout image data as a reference (100 [%]). Is generated by reducing and enlarging a plurality of frames. Taking the frame division image data of FIG. 7A as an example, as shown in FIG. 8, 10% from 10% to 90% by reducing the image size with respect to the reference frame division image data. Nine different frame size image data is generated in increments of [%]. Furthermore, by dividing the image size, frame divided image data of 10 different sizes is generated in increments of 10 [%] from 110 [%] to 200 [%]. As a result, frame divided image data having a size of 20 levels is obtained.
The frame division information extraction unit 11 compresses the frame division image data having a size of 20 levels by a compression method having reversibility such as run-length compression, and the compressed frame division image data is converted into comic information. The data are stored in the storage device 70 in association with each other (step S112). Note that such frame allocation information is generated for all comics to be detected.

Next, the operation of the frame division information extraction process in the comic image data detection apparatus 100 will be described.
Now, in the comic image data detection apparatus 100, it is assumed that the operator operates the input device 74 and inputs a patrol start instruction. Thereby, the website search unit 13 reads the search word included in the comic information stored in association with the comic image data in the storage device 70 in response to the input of the patrol start instruction (Yes in step S200) ( Step S202). In addition to patrol start instruction input by the operator, patrol start time is scheduled in advance by dedicated software, and Web site search processing is automatically executed when the scheduled start time is reached. You may comprise.

  Next, the website search unit 13 searches the website using the read search word (step S204). Here, the search of the Web site is performed by using, for example, an OR condition, using the title of the comic to be searched, the author name, and the number of comic volumes. That is, if a website includes a page including any one of these search words, the website is searched (hit). Moreover, as a search engine, for example, a well-known and reliable search engine is used.

  When a Web site corresponding to the search word is searched by this search processing (Yes in Step S206), the Web site search unit 13 uses the searched Web site information (URL) as comic information of the comic to be searched. And stored in the storage device 70 (step S208). Then, the image data acquisition unit 14 is notified that the Web site information has been stored.

  If the image data acquisition unit 14 determines that the notification from the Web site search unit 13 has been received, the image data acquisition unit 14 here performs the hit order (display order at the time of search) based on the Web site information (URL) stored in the storage device 70. Then, each website is accessed, and image data published on the accessed website is acquired (step S210). Then, the image data acquisition unit 14 stores the acquired image data in the storage device 70 in association with the comic information (step S212). Note that, as described above, a configuration may be adopted in which all image data that matches a preset acquisition condition is acquired from a Web site without searching the Web site using a search word. In this case, however, information (for example, URL) of the website that publishes image data that meets the acquisition conditions is stored in association with the acquired image data together with the image data.

The image data acquisition unit 14 determines whether or not the acquisition number of the image data has reached a preset upper limit number (step S214). Here, it is determined that it has not been reached (No in step S214), and it is determined that the acquisition of image data has been completed (step S238). Thereby, the image data acquisition unit 14 notifies the matching unit 15 that the acquisition of the image data is completed.
When the matching unit 15 determines that the notification from the image data acquisition unit 14 has been received, the frame allocation corresponding to the acquired image data (acquired image data) and the comic information associated with the acquired image data from the storage device 70. Image data is read out (step S216).

Then, the matching unit 15 executes a matching process between the acquired acquired image data and the frame-cut image data (step S218).
Specifically, the matching unit 15 first converts the read acquired image data into edge image data. At this time, the frame division information extraction unit 11 converts the image data into edge image data by a processing method similar to the processing method used during the frame division information extraction processing.

  Next, in accordance with the same rule as the extraction rule used in the frame division information extraction unit 11, the matching unit 15 leaves the portion estimated as the frame division line portion from this edge image data, and removes other drawing portions. Frame image data (acquired frame image data) is extracted. As described above, it is also possible to employ a configuration in which matching is performed using acquired image data as it is without performing conversion to edge image data.

  The matching unit 15 performs template matching on the extracted acquired frame divided image data using each frame divided image data read from the storage device 70 as template data. At this time, matching may be performed with all of the 20-stage frame-divided image data, but here, the acquired frame-divided image is detected for one page of comic and two pages of spread or disjoint. Depending on the size of the data, the same or the closest frame divided image data and half the frame divided image data are selected. Then, matching is performed using the selected frame division image data. Note that for two pages, matching is performed with all combinations of frame-cut image data corresponding to comics to be detected.

  The matching is performed by superimposing the acquired frame-cut image data and the frame-cut image data, and using the pixel values of each of the two pixels at the overlapping position. Here, the size (resolution) of both images is matched so that they overlap exactly. For the half-sized frame divided image data, the size is adjusted so that it overlaps exactly with one or the other half of the acquired frame divided image data. Then, for each pixel in the overlapping position when the two are overlapped with reference to the upper left or upper right pixel of the image, the absolute value of the difference between the pixel values is calculated, and the sum of the absolute values is further calculated. Is calculated. The matching unit 15 outputs the sum of the selected frame-cut image data as a matching result to the determination unit 16.

  When the determination unit 16 acquires the matching result from the matching unit 15, the determination unit 16 compares the acquired matching result with a preset matching threshold value. Then, when the sum is equal to or less than the matching threshold, it is determined that the acquired frame-cut image data matches or is similar to the frame-cut image data used for matching. On the other hand, when the sum is larger than the matching threshold, it is determined that the acquired frame-cut image data is not similar to the frame-cut image data used for matching (step S220). Here, in the above matching method, the smaller the sum is, the higher the similarity between the acquired frame-cut image data and the frame-cut image data is (the sum is “0”, which is a perfect match). The matching threshold value may be a single fixed value that is common to all, a threshold value may be set for each framed image data, or a threshold value may be set for each size of each framed image data. Also good. In particular, as the number of frame dividing lines increases, the degree of coincidence is predicted to decrease. Therefore, if a very small threshold is set for such frame dividing image data, the detection rate decreases. There is a risk of inviting. On the other hand, if the threshold is increased too much, it is determined that even unnecessary images are matched. Therefore, it is desirable to set an optimum threshold based on the results of tests performed in advance, measured data, and the like.

The determination unit 16 generates determination result information including the determination result information thus determined and the Web site information, and outputs the generated determination result information to the detection unit 17.
Based on the determination result information from the determination unit 16, the detection unit 17 determines whether the matched acquired image data includes similar image data (step S222). If it is determined that similar image data is included (Yes in step S222), the detection unit 17 adds 1 to 2 here to the number of similar image data of the corresponding website (step S224). . Specifically, 1 is added when the content of one page is included in one acquired image data, and 2 is added when the content of two pages is included.
The series of matching processes (steps S216 to S224) are repeatedly executed until the process is completed for all acquired image data.

  Then, it is determined that the matching process is completed in the matching unit 15 (Yes in Step S226), and it is determined in the image data acquisition unit 14 that there is no unacquired image data (No in Step S228). Thereby, the detection unit 17 compares the number of similar image data counted so far with a preset threshold value for each Web site. Then, based on the comparison result, it is determined whether or not there is a match with the narrow-down condition (step S232). Here, it is assumed that the count threshold is set to “3” to determine that it is an illegal candidate site when “3” or more pages of comics to be detected are included. Accordingly, when the detection unit 17 determines that the number of similar image data is “3” or more, the detection unit 17 determines that the filtering condition is met (Yes in step S232), and the number of similar image data is less than “3”. Is determined as not satisfying the narrow-down condition (No in step S232).

  If the detection unit 17 determines that the filtering condition is met, the information (URL) of the corresponding Web site (illegal candidate site) and at least acquired image data that is determined to be similar and illegal are detected. Detection result information in which a part is associated with comic information is generated (step S234). On the other hand, when it is determined that the filtering condition is not met, detection result information in which the information of the corresponding website is associated with the comic information is generated (step S236).

Then, the detection unit 17 stores the generated detection result information in the storage device 70 and notifies the detection result display unit 18 that the detection result information has been stored (or updated) (step S238). In addition, the detection unit 17 deletes unnecessary acquired image data including acquired image data corresponding to a Web site that is determined not to be illegal from the storage device 70.
The detection result display unit 18 reads the detection result information from the storage device 70 in response to the notification from the detection unit 17, and displays the detection result based on the read detection result information (step S240). When displaying the detection results, the detection results may be sorted and displayed in descending order of similarity.

As described above, in the comic image data detection apparatus 100 according to the present embodiment, the content of the comic image portion removed from the comic image data of the comic to be detected while leaving the frame dividing line drawing portion as the frame dividing information. Can be extracted. Furthermore, with the extracted frame-cut image data as a reference, a plurality of frame-cut image data of different sizes can be generated by reducing and enlarging the frame-cut image data in a range of 10 [%] to 200 [%]. . Furthermore, data obtained by compressing these frame-segmented image data by a compression method having reversibility such as run-length compression can be stored in the storage device 70 as frame-segmentation information.
The image of the frame dividing line portion forming each frame is a single color (black), and the frame dividing line is basically straight line information, so the drawing portion is less than the comic portion. Therefore, even if the frame division information is compressed by a reversible compression method, the data amount can be greatly reduced.

The frame division image data is data including information such as the shape, position, and size of each frame, and has a function as template data used for matching. In addition, since the frame allocation information is information unique to comics, the amount of data is greatly increased without reducing the detection accuracy compared to matching using an original image or an edge image including all edge components. Can be reduced.
In addition, since the comic image data detection apparatus 100 according to the present embodiment generates binarized image data as frame divided image data, for example, a difference from a pixel having a pixel value “0” of the frame divided image data. Can be omitted, and the time required for the matching process can be shortened as compared with matching using an original image (particularly, a multi-tone image).

Further, the comic image data detection apparatus 100 according to the present embodiment can narrow down and detect Web sites that may illegally publish comic image data depending on the narrowing conditions. It is possible to reduce the trouble of manually detecting the site.
Here, in the above embodiment, the comic image data acquisition unit 10 constitutes a comic image data acquisition unit, and the frame division information extraction unit 11 constitutes a frame division information extraction unit, and the frame division information extraction unit 11 and the storage The device 70 constitutes frame allocation information storage means.

In the above embodiment, the website search unit 13 and the image data acquisition unit 14 constitute an image data acquisition unit, and the matching unit 15 constitutes a matching unit.
Moreover, in the said embodiment, the determination part 16 and the detection part 17 comprise a detection means, and the detection part 17 and the memory | storage device 70 comprise a site information storage means.
Further, in the above embodiment, steps S108 to S110 correspond to the frame allocation information extraction step, and step S112 corresponds to the frame allocation information storage step.
In the above embodiment, steps S204 to S210 correspond to the image data acquisition step, step S218 corresponds to the matching step, and steps S220 to S236 correspond to the detection step.

(Modification)
In the above-described embodiment, the frame dividing information is extracted as the frame dividing information by leaving the frame dividing line portion and removing the other cartoon portions. However, the present invention is not limited to this configuration.
For example, in the comic, there is a balloon part that describes words such as a person in each frame as characteristic information other than the frame division. By leaving the drawing contents of the balloon part as well, it is possible to generate frame allocation information with higher detection accuracy. The drawing contents of the speech balloon part include a frame line part constituting the speech balloon and a character part described in the speech balloon. Either one of these may be extracted, or both may be extracted. The character portion may be left in the image. However, the character portion may be converted into a character code and stored separately in association with the position information of the balloon.

  Further, not only the balloon part but also a frame in which the face of the character is drawn may be stored as frame division information. However, as the amount of template data increases, the time required for the matching process becomes longer, so it is included in a trade-off with the processing time. Alternatively, in the first matching process, matching is performed using frame-cut image data in which only the drawing information of the frame dividing line portion is left, and only the image data of the website that is determined to be possibly disclosed illegally by this matching On the other hand, the matching process using the frame-cut image data including the position information of the balloon part and the face is performed. The information of the balloon part is held as difference data in consideration of the data storage capacity.

  In the above embodiment, the frame division image data is extracted as the frame division information. However, the present invention is not limited to this configuration. For example, if the shape, size, and position of each frame are understood, for example, if the shape of the frame is a polygonal shape, the coordinate information of the vertices of each frame, the information indicating the shape, and the size of the comic image data (page size) It is good also as a structure which extracts. In this case, for example, at the time of the matching process, the same information is extracted from the acquired image data and matching is performed, and the image data is reconstructed based on the information and matching with the reconstructed image data is performed. There is a configuration to perform.

In the above embodiment, the comic image data is acquired from the image scanner. However, the present invention is not limited to this configuration. For example, it may be configured to acquire via a compact memory such as a USB memory, a CD drive, a DVD drive, or a Blu-ray drive, or may be acquired from another PC via a network.
The above embodiments are preferable specific examples of the present invention, and various technically preferable limitations are given. However, the scope of the present invention is described in particular in the above description to limit the present invention. As long as there is no, it is not restricted to these forms. In the drawings used in the above description, for convenience of illustration, the vertical and horizontal scales of members or parts are schematic views different from actual ones.
Further, the present invention is not limited to the above-described embodiment, and modifications, improvements, and the like within the scope that can achieve the object of the present invention are included in the present invention.

  DESCRIPTION OF SYMBOLS 1 ... Comic image data detection system, 100 ... Comic image data detection apparatus, 200_1-200_N ... Web server, 10 ... Comic image data acquisition part, 11 ... Frame allocation information extraction part, 12 ... Data communication part, 13 ... Web site search , 14 ... Image data acquisition part, 15 ... Matching part, 16 ... Determination part, 17 ... Detection part, 18 ... Detection result display part, 60 ... CPU, 62 ... RAM, 64 ... ROM, 66 ... I / F, 68 ... Bus, 70 ... Storage device, 72 ... Display device, 74 ... Input device, 76 ... Image scanner, 78 ... NIC

Claims (7)

A comic image data detection device that detects image data that matches or is similar to comic image data, which is data obtained by converting a preset page in a comic to be detected into image data from image data existing on a website,
A frame division that extracts frame division information that is information related to a frame division including at least the shape of each divided frame and the position of each frame in the comic image data from the comic image data of the comic to be detected Information extraction means;
Frame allocation information storage means for storing frame allocation information extracted by the frame allocation information extraction means;
Image data acquisition means for acquiring image data from the website;
Based on the image data acquired by the image data acquisition unit and the frame allocation information stored in the frame allocation information storage unit, a matching unit that performs matching between the acquired image data and the frame allocation information;
Detection means for detecting image data that matches or is similar to the comic image data from the acquired image data based on the similarity between the acquired image data and the comic image data based on the matching result of the matching means; A comic image data detection device comprising:   The frame division information extracting means converts the comic image data into edge image data which is image data indicating a distribution of edge components included in the image of the comic image data, and the frame division information is converted from the edge image data. 2. The comic image data detection apparatus according to claim 1, wherein frame-composed image data, which is image data in which only edge components corresponding to the outer edge portion of each frame-decomposed frame are drawn, is extracted.   The frame allocation information extraction unit is configured to change the image size of the extracted frame allocation image data corresponding to the comic image data while changing the image size while maintaining the aspect ratio as the frame allocation information. The comic image data detection apparatus according to claim 2, wherein a plurality of different frame division image data are generated. The detection means determines, for each Web site, whether or not the detected image data matches a preset narrowing condition based on the image data detected as the matching or similar image data, If it matches, the image data is detected as comic image data that may have been illegally disclosed.
It comprises site information storage means for storing comic image data detected by the detection means, which may be illegally disclosed, and information on a website that discloses the image data in association with each other. The comic image data detection apparatus according to any one of claims 1 to 3.   The detection means determines that the filtering condition is met when the image data detected as the matching or similar image data includes comic image data of a preset number of pages or more. The comic image data detection device according to claim 4. A comic image data detection program for detecting image data matching or similar to comic image data, which is data obtained by converting a preset page of a comic to be detected into image data from image data existing on a website,
A frame division that extracts frame division information that is information related to a frame division including at least the shape of each divided frame and the position of each frame in the comic image data from the comic image data of the comic to be detected Information extraction means,
Image data acquisition means for acquiring image data from the website;
Based on the image data acquired by the image data acquisition unit and the frame allocation information stored in the frame allocation information storage unit in which the frame allocation information extracted by the frame allocation information extraction unit is stored, A matching means for matching with the frame allocation information, and
Based on the degree of similarity between the acquired image data and the comic image data based on the matching result of the matching means, it is realized as a detection means for detecting image data that matches or is similar to the comic image data from the acquired image data. A comic image data detection program characterized by including a program for causing a computer to execute a function to be executed. A comic image data detection method for detecting image data matching or similar to comic image data, which is data obtained by converting a preset page in a comic to be detected into image data from image data existing on a website,
A frame division that extracts frame division information that is information related to a frame division including at least the shape of each divided frame and the position of each frame in the comic image data from the comic image data of the comic to be detected An information extraction step;
Frame allocation information storage means for storing the frame allocation information extracted in the frame allocation information extraction step;
An image data acquisition step of acquiring image data from the website;
Based on the image data acquired in the image data acquisition step and the frame allocation information stored in the frame allocation information storage step, a matching step for matching the acquired image data and the frame allocation information;
A detection step of detecting image data matching or similar to the comic image data from the acquired image data based on a similarity between the acquired image data and the comic image data based on a matching result of the matching step; A comic image data detection method comprising:

Images