JP5320007B2 - Information filter device - Google Patents

Description

  The present invention relates to an information filter device, and more particularly to an information filter device that regulates the display of harmful information on a Web page.

Currently, services that prohibit browsing of specific Web pages for minors and the like are being performed on mobile phones and computers. Such service processing is called harmful information filtering (hereinafter simply referred to as filtering). Two types of filtering, a white list method and a black list method, are known.
In the white list method, a list of websites that can be browsed is listed, and browsing is uniformly permitted for pages included in the websites listed. In the black list method, a list of Web sites that are prohibited from browsing is listed, and browsing is uniformly restricted for pages included in the Web sites listed.
"Blacklist? Whitelist? Basic knowledge for" Keitai filtering principle "", Nikkei Trendy Net, 2007, December 26, [Search September 26, 2008], Internet, <URL: http: / /trendy.nikkeibp.co.jp/article/special/20071226/1005687/>

However, in the above-described white list method and black list method, browsing of the entire website is restricted only by a part of the web page of the website containing harmful information. For this reason, the user cannot obtain information that is in a Web site partially including harmful information and does not need to be restricted in browsing.
In addition, there is a problem for Web site operators in that it is costly to reduce the number of viewers and to monitor harmful information on their Web site.
The present invention has been made in view of the above points, and restricts browsing for each piece of information contained in a web page in a website, and information that does not need to be regulated even for a website that contains harmful information. An object of the present invention is to provide a filter device that enables browsing.

In order to solve the above problems, an information filter device according to claim 1 of the present invention acquires display data including content data representing at least one of text and an image and control information for displaying the content data. Display data acquisition means, content extraction means for extracting only content data from the display data, outline detection means for detecting the proportion of lines in content data with respect to content data representing an image, and the content extraction means The extracted content data is collated with preset harmful information, the harmful information collating means for judging whether the content data and the harmful information match, and the harm information as a result of the collation by the harmful information collating means. Content data that is determined to be consistent with other information that is not harmful information And a content replacement means for replacing the, the content extraction means, the extracted content data, classified according to attributes of the text or images, the harmful information collating means, the categorized content data, the attribute When the proportion of the line in the content data representing the image detected by the contour detection means is larger than a predetermined threshold, the content data representing the image and the harmful information are matched. A display in which the content data replaced by the content replacement unit is integrated into the display data, and only harmful information is replaced with other information that is not harmful information, out of the content data included in the display data, by omitting the verification process It is characterized by generating data. According to such an invention, among the contents included in the Web page, only the display of harmful information can be restricted, and the content that does not correspond to the harmful information included in the same Web page can be displayed. For this reason, the user can obtain information that does not require browsing even if the web page partially includes harmful information. Moreover, even if harmful information is written on a part of the Web page, browsing of the entire Web site is not restricted, so the burden of monitoring harmful information on the Web site operator is reduced. can do. In addition, according to such an invention, since only harmful information that is highly likely to match content data can be targeted for verification, the time required for verification and the processing load can be reduced. Furthermore, according to such an invention, content data that has a lot of lines and is difficult to identify can be omitted from the processing target for matching with harmful information, so the time required for matching while preventing browsing of harmful information. And the processing load can be reduced.

Hereinafter, an embodiment of a filter device according to the present invention will be described with reference to the drawings.
(Device configuration)
FIG. 1 is a functional block diagram for explaining the filter device of the present embodiment. In the present embodiment, the filter device is configured as the filter processing unit 109 of the mobile device 102. The illustrated filter processing unit 109 is mounted on the mobile device 102 and communicates with an NW (Net Work) device 101. The illustrated NW device 101 indicates the entire device related to the NW including the server in which the Web page is stored. The mobile device 102 is a mobile communication terminal device such as a mobile phone.

  The mobile device 102 includes a signal transmission / reception unit 110 that receives data from the NW device 101. The signal transmission / reception unit 110 also receives HTML (Hyper Text Markup Language: description language for displaying a Web page) data (display data in this embodiment) downloaded from the NW device 101 (hereinafter referred to as DL). To do. In addition, the mobile device 102 includes a display device unit 111 that displays DL data, and a filter processing unit 109 that regulates information displayed on the display device unit 111.

  The filter processing unit 109 includes an information extraction unit 103 that extracts content data (hereinafter referred to as “content”) from HTML data received by the signal transmission / reception unit 110, and information that is defined as harmful (hereinafter referred to as “harmful information”). Information storage unit 107 for storing information, information collation unit 104 for collating the extracted content with harmful information stored in information storage unit 107 to detect harmful information from the content, extracted harmful information Information replacement unit 106 that replaces the part with other information that is not harmful information, and creates display data that is displayed by embedding content including information obtained by replacing harmful information (hereinafter referred to as replacement information) in the received HTML data The information integration unit 105 is provided. The display data created by the information integration unit 105 is sent to the display device unit 111 and displayed.

In the above configuration, the information extraction unit 103 divides HTML data into an HTML tag unit and content. The HTML tag portion is control information that can instruct the display software such as a Web browser on the document structure, format, character decoration, etc., or embed an image or a link to another document.
Also, the mobile device 102 DLs a database (DB) of harmful information from the NW device 101 in advance and receives it by the signal transmission / reception unit 110. The received DB of harmful information is stored in the information storage unit 107.

In this specification, the term “content” refers to image data representing an image or text data representing a text, and refers to a predetermined unit of data. Content representing text is further classified into word data and document data. In the present embodiment, the types of image data, word data, and document data are hereinafter also referred to as “attributes”.
The predetermined unit can be set, for example, as one word for word data, a paragraph composed of a plurality of documents for document data, and one image for image data.
Needless to say, the present embodiment does not limit the content to such a unit, but can arbitrarily set the content.

(Harmful information DB)
FIG. 2 is a diagram for explaining a specific method in which the information storage unit 107 stores a DB of harmful information. As shown in the figure, harmful information is classified according to the attributes of word data, document data, and image data, and stored with a management number.
The information collation unit 104 collates only the content extracted by the information extraction unit 103 with harmful information stored in the information storage unit 107. As a result, it is determined that the content determined to match the harmful information cannot be displayed on the display device unit 111. Further, the information matching unit 104 determines that the content determined not to match the harmful information is content that can be displayed on the display device unit 111.

As will be described later, the information matching unit 104 also has a function of detecting the proportion of lines in the image content.
Among the configurations described above, the signal transmission / reception unit 110 functions as a display data acquisition unit. The information extracting unit 103 functions as a content extracting unit, the information matching unit 104 functions as a harmful information matching unit and an outline detecting unit, and the information replacing unit 106 functions as a content replacing unit.

(Verification of content and harmful information)
FIG. 3 is a diagram for explaining a method of collating content with harmful information in the information collating unit 104 shown in FIG. Hereinafter, a method for collating content with harmful information will be described in detail for each category of content word data, document data, and image data.
-Word data and document data As the collation processing of word data and document data, the extracted word data and document data contents are compared character by word with harmful information in word data and document data. As a result of the comparison, the information matching unit 104 determines that the content is harmful information not only when the characters of the content and harmful information completely match, but also when the content includes characteristic characters of the harmful information. Judge that there is.

That is, for example, as shown in FIG. 3A, it is assumed that harmful information of word data is “ABCDE”. Further, the contents collated with the harmful information are the contents 1 “ABCDE” shown in FIG. 3B, the contents 2 “AB × CΔD □ E” shown in FIG. 3C, and FIG. Content 3 “ACDA” shown in FIG.
In such a case, contents including other information as well as all harmful information, such as content 1 and content 2, are determined to be harmful information. In addition, content that does not include a part of characters that constitute harmful information, such as content 3, is not determined to be harmful information.

In addition, in the case of image data content, pattern information is collated with harmful information of image data stored in the information storage unit 107. In pattern matching, for example, the correlation between the two images is taken, the height of the correlation is compared with a preset threshold value, and the content matches the harmful information when the correlation exceeds the threshold value. To be judged. Such pattern matching methods include, for example, the JPO homepage, image recognition technology, search on September 26, 2008, the Internet,
<URL: http://www.jpo.go.jp/shiryou/s_sonota/tokumapf.htm>
Since it is well known, no further explanation will be given. Note that the present embodiment is not limited to determining whether or not the content is harmful information as described above, and may be determined by any method.

  In the present embodiment, contents are classified according to attributes, and harmful information is classified and stored for each attribute of word data, document data, and image data as shown in FIG. For this reason, when content is collated, only harmful information that matches content attributes can be targeted for collation, and the time required for collation processing can be shortened.

In the present embodiment, the time required for pattern matching can be shortened by performing pattern matching after reducing the content and harmful information. In addition, such a method is, for example, Hiroshi Nakajima, “Faster pattern matching method using FPGA”, search on September 26, 2008, the Internet,
As described in <URL: http: //www.hfl.hiroshima-u.ac.jp/COE/programs/pdf/nakano.pdf>, since it is publicly known, no further explanation will be given.
Furthermore, in this embodiment, prior to pattern matching, the information matching unit 104 may detect the proportion of lines in the content of image data. In this way, when the line ratio exceeds the threshold value, it is possible to reduce the processing time by omitting pattern matching because the content is difficult to identify.

  The content collated with the harmful information as described above is sent to the information replacement unit 106 as both content that can be displayed and content that cannot be displayed. That is, the content that is determined to be harmful information by the information matching unit 104 and the content that is determined not to be harmful information are input to the information replacing unit 106. The information replacement unit 106 replaces only the content determined to be harmful information with other information that is not harmful information. When the other information is a character string such as “cannot be displayed”, for example, the character string “cannot be displayed” is set to be displayed in the content portion determined to be harmful information. When the other information is so-called “painting”, the content portion determined to be harmful information is set to be displayed in black.

  The information integration unit 105 performs integration processing using the HTML data received by the signal transmission / reception unit 110 and the content replaced or not replaced by the information replacement unit 106. In the integration process, the replaced content is replaced with the content before replacement in the HTML data, and is embedded in the HTML data. Then, the HTML data in which the replaced content is embedded is reproduced again. The reproduced HTML data is input to the display unit 111 and displayed. As a result, the display unit 111 displays HTML data in which only the harmful information portion is prohibited from being viewed.

(Operation)
Next, the operation of the filter device of the present embodiment having the above-described configuration will be described.
FIG. 4 is a timing chart for explaining a communication procedure performed between the NW device 101 and the mobile device 102 when browsing a Web page. In the mobile device 102, the browsing control unit 401 starts and ends browsing of the Web page. In addition, the start and end of browsing of the Web page are transmitted from the browsing control unit 401 to the filter processing unit 109.

The signal transmission / reception unit 110 of the mobile device 102 requests the NW device 101 to access the Web page (S1). At this time, in the example shown in FIG. 4, the mobile device 102 DLs the DB of harmful information from the NW device 101 as needed (S2, S3). In such an example, every time the mobile device 102 DLs a Web page from the NW device 101, the latest DB such as harmful information is DL. The received harmful information and the like are temporarily stored in the information storage unit 107 in the mobile device 102 (S4).
When the Web page is downloaded from the NW device 101 (S5), browsing of the Web page is started in the mobile device 102 (S6). When the browsing of the web page is completed (S8), the harmful information stored in the web page and the information storage unit 107 is released (S7, S9).

The DL of the harmful information DB is not limited to such an example, and may be performed at any timing as long as the DL is performed before the filtering process. Then, the harmful information that has been DL may be stored in the information storage unit 107 in a nonvolatile manner. In such a case, certain harmful information or the like is stored in the information storage unit 107 when the mobile device 102 is shipped. Thereafter, each time the harmful information DB is updated in the NW device 101, the information storage unit 107 may be updated using ADL (air down load: a method of updating software or the like through a wireless interface) or the like.
The NW device 101 that has received the access request to the Web page DLs the requested Web page to the mobile device 102 (S5). In the mobile device 102, DL Web page data is handled in the HTML data format.

  FIG. 5 is a flowchart for explaining processing performed on a Web page that is downloaded by the filter processing unit 109 according to the embodiment of this invention. When the HTML data is received by the signal transmission / reception unit 110 (step S501), in the filter processing unit 109, the information extraction unit 103 shown in FIG. 1 extracts only the content from the HTML data. That is, the HTML data includes content and an HTML tag part. The filter processing unit 109 extracts only the content by dividing the HTML data into the content and the HTML tag unit (step S502). Then, the extracted content is classified for each attribute.

  Next, in the filter processing unit 109, the information collation unit 104 shown in FIG. 1 collates the classified content with the harmful information stored in the information storage unit 107, and determines whether or not the content is harmful information. Judgment is made (step S503). As a result of the determination, if the content is harmful information (step S503: harmful information), the information replacement unit 106 shown in FIG. 1 converts the content determined to be harmful information into a text such as “cannot be displayed”. Replace (step S504). The information integration unit 105 shown in FIG. 1 embeds and integrates the replaced content by replacing it with the content before replacement in the HTML data (step S505).

If it is determined in step S503 that the content is not harmful information (step S503: not harmful information), the content and HTML data are integrated without replacing the content. The integrated data is input to the display device unit 111 as display data (step S506). The display device unit 111 shown in FIG. 1 displays display data on a display screen (not shown).
When browsing of the display data is completed, the user closes and releases the Web page. Information indicating the release is sent to the signal transmission / reception unit 110 (S8), and the mobile device 102 issues a Web page release request to the NW device 101 (S9). Information indicating the release is also transmitted to the filter processing unit 109 shown in FIG. 1 (S7).

(system)
FIG. 6 is a diagram for explaining a system constructed by the mobile device 102 and the NW device 101 to which the filter device of this embodiment is applied. In the system of FIG. 6, when the mobile device 102 includes the filter processing unit 109 and the user requests the NW device 101 to browse the website from the mobile device 102 (in the figure: Web page request), The NW device 101 downloads the web page of the web site requested to the mobile device 102 (in the figure: Web page DL).

  As described above, the mobile device 102 extracts the content from the HTML data of the DL Web page, compares the harmful information stored in the information storage unit 107 in advance with the extracted content, and the two match. Judge whether or not. Then, only content that matches the harmful information is replaced with other data that is not harmful information such as “cannot be displayed”. As a result, the display unit 111 displays a Web page in which only words, documents, or images determined to match the harmful information in the HTML data are not displayed.

(Examples of other systems)
FIG. 7 is a diagram for explaining another example of the system of the present embodiment. In the system illustrated in FIG. 7, the NW device 101 includes the filter processing unit 109. In such a case, whenever the mobile device 102 requests the NW device 101 to browse the web page, the web page requested to be browsed in the NW device 101 is acquired by the filter processing unit 109 and filtered, and harmful information is obtained. The Web page from which the message is excluded is transmitted to the mobile device 102.
In the configuration illustrated in FIG. 7, the filter processing unit 109 acquires the HTML data of the Web page directly from the NW device 101 instead of the signal transmission / reception unit 110 illustrated in FIG. 1. For this reason, the display data acquisition means in the configuration of FIG.

Furthermore, the present embodiment provides a display data acquisition function for acquiring display data including content representing at least one of text and images and control information for displaying the content, and content for extracting only the content from the display data. The extraction function, the extracted content function and harmful information are collated, the harmful information matching function that determines whether the content and harmful information match or not, and the content that is determined to match the harmful information as a result of matching, The configuration similar to that of the above-described embodiment can be realized by DL, on a general-purpose computer, a program that causes a computer to execute a content replacement function that replaces other information that is not harmful information. When using a DL program, it is possible to allow only a specific user to have the authority to set and cancel by using authentication such as a password. Further, by associating the above program with a UIM (User Interface Module) card, it is possible to distinguish between setting and canceling by the UIM even in the same terminal.
According to such a system, the filtering process can be executed in any of the communication terminal device and the network device. For this reason, it is possible to execute the filter processing on a preferable side according to the use of the system and the target of application.

It is a functional block diagram for demonstrating the filter apparatus of one Embodiment of this invention. It is a figure for demonstrating the specific method in which the information storage part shown in FIG. 1 stores DB of harmful information. It is a figure for demonstrating the method of collation with the content and harmful information in the information collation part shown in FIG. It is a timing chart for demonstrating the procedure of the communication performed between the NW apparatus of one Embodiment of this invention, and a mobile apparatus. It is a flowchart for demonstrating the process performed with respect to the Web page DLd in the filter process part of one Embodiment of this invention. It is a figure for demonstrating the system constructed | assembled by the mobile apparatus and NW apparatus with which the filter apparatus of one Embodiment of this invention is applied. It is a figure for demonstrating the other example of the system of one Embodiment of this invention.

Explanation of symbols

101 NW device 102 Mobile device 103 Information extraction unit 104 Information collation unit 105 Information integration unit 106 Information replacement unit 107 Information storage unit 109 Filter processing unit 110 Signal transmission / reception unit 111 Display device unit 401 View control unit

Claims (1)

Display data acquisition means for acquiring display data including content data representing at least one of text and image and control information for displaying the content data;
Content extraction means for extracting only content data from the display data;
Contour line detecting means for detecting a ratio of the line in the content data to the content data representing the image;
Harmful information collating means for collating the content data extracted by the content extracting means with preset harmful information and determining whether the content data and harmful information match or not;
Content replacement means for replacing content data determined to match the harmful information as a result of matching by the harmful information matching means with other information that is not harmful information;
With
The content extraction means classifies the extracted content data for each attribute of text or image,
The harmful information collating means collates the classified content data only with harmful information having matching attributes,
When the proportion of the line data in the content data representing the image detected by the contour detection means is greater than a predetermined threshold, the process of matching the content data representing the image and harmful information is omitted,
Integrating the content data replaced by the content replacement means into the display data, and generating display data in which only harmful information of the content data included in the display data is replaced with other information that is not harmful information. A characteristic information filter device.

Images