WO2015099303A1

WO2015099303A1 - System for analyzing objectionable image

Info

Publication number: WO2015099303A1
Application number: PCT/KR2014/011535
Authority: WO
Inventors: 김유신
Original assignee: (주)티디엘
Priority date: 2013-12-23
Filing date: 2014-11-28
Publication date: 2015-07-02
Also published as: KR20150073429A

Abstract

Disclosed is a system for analyzing an objectionable image. The system for analyzing an objectionable image according to one embodiment of the present invention comprises: an analysis server for analyzing whether an image is objectionable or not by mixing a hash and DB-based blocking scheme and a modified contents-based blocking scheme which extracts characteristic points in a region of interest (ROI) of a human body; and a web server for, upon an online image upload request from a contents register terminal or an online reproduction request for an image from a user terminal, requesting the analysis server to analyze whether the image to be uploaded or the image to be reproduced is objectionable or not so as to control the uploading of the image to be uploaded or the reproducing of the image to be reproduced according to the analysis result.

Description

Hazardous Image Analysis System

The present invention relates to a harmful image analysis system, and more particularly, a harmful image is detected by using a hash and DB-based blocking technique and a modified content-based blocking technique that extracts a feature point only in the ROI of an image. The present invention relates to a technology for controlling playback and blocking of uploaded or harmful images on-line.

The Internet is not only a variety of information that is called the sea of information, but it is also easy to use, making it a part of everyday life for many people in modern times and providing positive aspects in social, economic and academic aspects. However, in contrast to this positive aspect, the indiscriminate dissemination of harmful information using the characteristics of openness, interconnection and anonymity of the Internet has become a serious social problem. This environment can attract emotionally and mentally harming young people who have poor value judgment and weak self-control. Therefore, there is a need for a method of blocking harmful information so that socially disadvantaged youth or unwanted people are not exposed to harmful information.

Conventionally, harmful image blocking methods include metadata and text information based blocking techniques, hash and DB based blocking techniques, and content based blocking techniques. The blocking technique based on metadata and text information is a method of determining the harmfulness of an image by analyzing the harmfulness of the text included in the title, file name, and description of the image, and has a high over-blocking rate and an error-blocking rate. Hash and DB based blocking method calculates database of hash value of known harmful image, calculates hash value of image and compares it with previously constructed database in order to judge the harmfulness of new image as input. This is a method of determining the harmfulness of the video. This method has a problem in that as the harmful image increases, the size of the hash value database increases as well as the amount of calculation for determining the harmfulness of the image increases. In addition, even if known harmful image has a problem that can not block if the hash value changes through a slight modification.

Recently, the content-based blocking technique has been analyzed to extract a feature by analyzing the content of the harmful image, generate a hazard classification model from the feature, and determine the harmfulness of the input image based on the generated hazard classification model. This method solves the high over-blocking rate and error-blocking rate of metadata and text-based blocking, and solves the database size and computational problems of hash and DB-based blocking.

However, most content-based blocking techniques use low-level features such as color, texture, and shape as the features of harmful images, or mainly MPEG-descriptors, which are mainly used for image retrieval. . This information does not properly reflect the characteristics of the harmful image, which results in low blocking rate and high error rate.

In order to solve this problem, a recent new approach detects skin color in pixels and utilizes skin and non-skin color ratios in the image as a characteristic of hazard determination. However, the approach using this feature is a technology that detects the human body part in the entire image and has a problem that it can be judged as a harmful image when the skin color occupies a high proportion even when it is not a harmful image.

[Preceding technical literature]

[Patent Documents]

Prior art related to the present invention is Korean Patent Publication No. 10-2010-0089349 (published: August 12, 2010).

Hash and DB-based blocking techniques and modified content-based blocking techniques that extract feature points only in the ROI of the image can be used to detect faded images and prevent harmful images from being uploaded online or playing and blocking playback. A harmful image analysis system for controlling is proposed.

The problem to be solved of the present invention is not limited to the above-mentioned problem, and other problems not mentioned may be clearly understood by those skilled in the art from the following description.

Harmful image analysis system according to an aspect of the present invention, an analysis server for analyzing the harmfulness of the image by using a hash and DB-based blocking technique and a modified content-based blocking technique to extract the feature point of the ROI; And according to the video upload request from the content registrant terminal or the playback request for the online video from the user terminal, analyzing whether the uploaded video or the playback requested video is harmful. And a web server configured to control the upload of the uploaded requested video or the playback of the requested video and the blocking of the playback according to the analysis result.

The harmful image analysis system further includes a database storing at least one fashionable image information, wherein the analysis server, the upload according to the analysis request for the harmfulness of the uploaded video or the playback requested video, The database is searched to determine whether the requested image or the reproduced requested image is a harmful image, and when it is determined that the image does not correspond to the harmful image, the hash and DB based blocking technique and the region of interest of the human body ( It is possible to analyze the presence of harmful images by using a modified content-based blocking technique that extracts the feature points of the ROI.

When the analysis server analyzes whether a harmful image is mixed by using the hash and DB based blocking technique and the modified content based blocking technique, the analysis server analyzes whether the harmful image is applied first by applying the hash and DB based blocking technique. If the harmful image is not analyzed, the modified content-based blocking technique may be applied to analyze the harmful image.

The analysis server constructs a hash value of at least one harmful image into a database for applying the hash and DB-based blocking schemes, and a hash value of the upload requested image or the playback requested image is built into the database. The harmfulness may be analyzed by determining whether it corresponds to a hash value of at least one harmful image.

According to another aspect of the present invention, a harmful image analysis system includes a hash and DB-based blocking technique and a human body for determining whether an image is harmful based on whether the hash value of the image is included in a database constructed as the hash value of at least one harmful image. An analysis server for analyzing the harmfulness of the image by using a modified content-based blocking technique for extracting feature points of the ROI; And a hash value of the image and a feature point of a region of interest (ROI) are calculated and transmitted to the analysis server to request an analysis of the harmfulness of the requested video, according to an analysis result. It includes a user terminal for controlling the playback and interruption of the image.

The analysis server may analyze the harmful image by applying the hash and DB-based blocking technique first, and analyze the harmful image by applying the modified content-based blocking technique when the harmful image is not analyzed. have.

According to the harmful image analysis system according to an embodiment of the present invention, a harmful image is detected by using a hash and DB-based blocking technique and a modified content-based blocking technique that extracts feature points only in the ROI of the image. By controlling the playback and blocking of harmful images uploaded online or by harmful videos, socially disadvantaged youth or unwanted people may not be exposed to harmful information.

1 is a view showing the configuration of a harmful image analysis system according to an embodiment of the present invention.

2 is a view showing the configuration of a harmful image analysis system according to another embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Embodiments of the present invention are provided to more fully explain the present invention to those skilled in the art, and the following embodiments may be modified in many different forms, the scope of the present invention It is not limited to the following embodiments. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the singular forms "a", "an" and "the" may include the plural forms as well, unless the context clearly indicates otherwise. Also, as used herein, "comprise" and / or "comprising" specifies the presence of the mentioned shapes, numbers, steps, actions, members, elements and / or groups of these. It is not intended to exclude the presence or the addition of one or more other shapes, numbers, acts, members, elements and / or groups. As used herein, the term "and / or" includes any and all combinations of one or more of the listed items.

Although the terms first, second, etc. are used herein to describe various members, regions, and / or portions, it is obvious that these members, components, regions, layers, and / or portions should not be limited by these terms. Do. These terms do not imply any particular order, up or down, or superiority, and are only used to distinguish one member, region or region from another member, region or region. Accordingly, the first member, region, or region described below may refer to the second member, region, or region without departing from the teachings of the present invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings schematically showing embodiments of the present invention. In the drawings, for example, variations in the shape shown may be expected, depending on manufacturing techniques and / or tolerances. Accordingly, embodiments of the present invention should not be construed as limited to the specific shapes of the regions shown herein, but should include, for example, changes in shape resulting from manufacturing.

Referring to Figure 1, the harmful image analysis system according to an embodiment of the present invention includes an analysis server 20 and the web server 10. At this time, the analysis server 20 and the web server 10 may be connected to the Internet and the network to perform data communication with each other to perform data communication. Accordingly, the analysis server 20 and the web server 10 should each be provided with a communication module (not shown) connected to the network to perform data communication. Meanwhile, the content registrant terminal 1 and the user terminal 2 should be provided with a communication module capable of performing data communication by being connected to a network such as the Internet. The content registrant terminal 1 is shown as a notebook in FIG. 1, but is not limited thereto. Any content terminal may be a terminal capable of performing data communication through a network connection such as the Internet. Similarly, the user terminal 2 is shown as a personal computer in FIG. 1, but is not limited thereto. Any user terminal 2 may be connected to a network such as the Internet to perform data communication.

The analysis server 20 analyzes whether the image is harmful by using a hash and DB-based blocking technique and a modified content-based blocking technique that extracts feature points of the ROI.

At this time, hash and DB based blocking technique calculates hash value of harmful image in advance and makes database, then calculates hash value of new image and compares it with previously constructed database in order to judge the harmfulness of new image as input. It is a method of determining the harmfulness of a new video. Accordingly, the analysis server 10 constructs a hash value of at least one harmful image as a database for applying the hash and DB-based blocking schemes, and the hash value of the upload requested image or the playback requested image is the database. The harmfulness may be analyzed by determining whether it corresponds to a hash value of at least one harmful image constructed as.

However, the harmful image determination method based on the hash and DB-based blocking technique has a problem that even if known harmful image can be blocked when the hash value is changed through a slight modification.

Therefore, to solve this problem, the analysis server 20 analyzes the harmfulness of the new image by using a hash and DB-based blocking technique and a modified content-based blocking technique that extracts the feature points of the ROI. . At this time, the modified content-based blocking technique for extracting the feature points of the ROI extracts the feature points of the ROI, for example, the face, the chest, the penis, etc. from the images, and generates a harmful classification model from the features. Then, based on the parasitic classification model generated, it is possible to determine the harmfulness of the incoming image.It solves the high over-blocking rate and error-blocking rate in metadata and text-based blocking, and the database size of hash and DB-based blocking. And a method of solving a computational problem, and extracting a feature point only to a region of interest of the human body, thereby solving a problem that may be determined as a harmful image when the skin color occupies a high proportion even when the image is not a harmful image.

As a result, the analysis server 20 is a modified content-based blocking technique for extracting feature points of the hash analyzer 21 and the ROI to analyze the harmfulness of a new image by applying a hash and DB-based blocking technique. The feature point analyzer 22 may analyze the harmfulness of the new image.

The web server 10 may request the upload of the uploaded video or the playback request in response to a video upload request from the content registrant terminal 1 online or a playback request for the video online from the user terminal 2. By requesting the analysis server 20 to analyze whether the video is harmful, it is possible to control the upload of the uploaded video or the play and block of the playback of the requested video according to the analysis result.

That is, in the case of a harmful image, the uploaded video is blocked from uploading or the reproduction of the playback requested video is blocked. On the other hand, if the image is not harmful, the uploaded video is uploaded online or the video requested to be played is a user. It can be controlled to play in the terminal (2).

On the other hand, the harmful image analysis system according to an embodiment of the present invention may further include a database 30 for storing at least one fashion image information.

Accordingly, the analysis server 20 determines whether the upload requested video or the playback requested video is a harmful video according to an analysis request regarding whether the uploaded video or the playback requested video is harmful. Search and identify the data, and use the modified content-based blocking technique that extracts the hash and DB-based blocking technique and the ROI's feature points when it is determined that the image does not correspond to the harmful image. It can analyze the presence of harmful images.

In this case, when analyzing the harmful image by mixing the hash and the DB-based blocking technique and the modified content-based blocking technique, the analysis server 20 analyzes the harmful image by first applying the hash and the DB-based blocking technique. If the harmful image is not analyzed through the modified content-based blocking technique, the harmful image may be analyzed. However, such an analysis sequence is only an example and is not limited thereto. That is, the analysis server 20 performs a modified content-based blocking technique that extracts the feature points of the ROI by changing the order first, and then analyzes the harmfulness by analyzing the harmfulness. The technique can be applied to analyze the harmfulness of the image. Alternatively, the analysis server 20 may analyze the harmfulness of the image by simultaneously performing a hash and DB-based blocking technique and a modified content-based blocking technique extracting feature points of the ROI.

2, the harmful image analysis system according to another aspect of the present invention, the user terminal 50 and the analysis server 60, the user terminal 50 and the analysis server 60, such as the Internet It is connected through a network to perform data communication, and for this purpose, a communication module (not shown) must be provided.

The analysis server 60 is a hash and DB-based blocking technique and a human interest region that determine whether the image is harmful by whether the hash value of the image is included in a database (not shown) constructed as the hash value of at least one harmful image. We analyze the harmfulness of images by using modified content-based blocking techniques that extract feature points of ROI. To this end, the analysis server 60 applies a hash and DB-based blocking technique, and a modified content-based blocking technique for extracting feature points of the hash analyzer 61 and ROI, which analyzes the harmfulness of the image. It may include a feature point analyzer 62 for analyzing whether the image is harmful by applying a.

In this case, the analysis server 60 analyzes the harmful image by applying the hash and DB-based blocking technique first, and analyzes the harmful image by applying the modified content-based blocking technique when the harmful image is not analyzed. can do.

However, this is only an example and is not limited thereto. That is, the modified content-based blocking technique may be applied first to analyze whether the harmful image is present, and if the harmful image is not analyzed through the hash and DB-based blocking technique, the harmful image may be analyzed. Alternatively, the harmful image may be analyzed by simultaneously applying the hash and the DB-based blocking technique and the modified content-based blocking technique.

The user terminal 50 calculates the hash value of the image and the feature points of the region of interest (ROI) according to the request of the reproduction of the image and transmits it to the analysis server 60 to request and analyze the harmfulness of the requested image. According to the result, the reproduction control and the reproduction blocking are performed. To this end, the user terminal 50 calculates the hash value of the image and the feature points of the region of interest (ROI) according to the request for playing the image and transmits the calculated points to the analysis server 60 to request an analysis of the harmfulness of the requested image. The playback unit 51 includes a playback control unit 53 for controlling the playback and playback interruption of the playback-requested video according to the system agent 52 and the analysis result of the analysis server 60. At this time, the playback control unit 53 prohibits the video of the video that is requested to be reproduced when it is analyzed as a harmful image as a result of analysis by the analysis server 60, and when it is analyzed that the video is not harmful, the video that is requested to be reproduced. Play through the screen (not shown) of the user terminal 50.

So far, the present invention has been described with reference to the embodiments. Those skilled in the art will understand that the present invention may be implemented in a modified form without departing from the essential characteristics of the present invention.

Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. Therefore, the scope of the present invention should not be construed as being limited to the above-described examples, but should be construed to include various embodiments within the scope and equivalents of the claims.

The present invention can be applied to the field of blocking u-hat in receiving content through a device that receives the content through the Internet, such as a PC, smart phone.

Claims

An analysis server for analyzing the harmfulness of the image by using a hash and DB-based blocking technique and a modified content-based blocking technique extracting feature points of the ROI; Wow

In response to a video upload request online from a content registrant terminal or a playback request for an online video from a user terminal, an analysis of the uploaded video or the harmfulness of the video requested to play is performed to the analysis server. And a web server for controlling the upload of the requested video or the reproduction of the requested video, and the blocking of the playback according to the analysis result.
The method according to claim 1,

The harmful image analysis system,

Further comprising a database for storing at least one fashion image information,

The analysis server may search the database to determine whether the upload requested video or the playback requested video is a harmful video according to an analysis request regarding whether the uploaded video or the playback requested video is harmful, If it is determined that the image does not correspond to the harmful image, the harmful image is analyzed by using the hash- and DB-based blocking technique and the modified content-based blocking technique that extracts the feature points of the ROI. Harmful image analysis system, characterized in that.
The method according to claim 1 or 2,

The analysis server,

When analyzing the harmful image by mixing the hash and DB based blocking technique and the modified content based blocking technique, the harmful image is analyzed by applying the hash and DB based blocking technique first and analyzing the harmful image. If not, the harmful image analysis system characterized in that for analyzing the harmful image by applying the modified content-based blocking technique.
The method according to claim 1,

The analysis server,

In order to apply the hash and DB-based blocking scheme, a hash value of at least one harmful image is being built into a database, and at least one harmful value of which the hash value of the upload requested image or the reproduction requested image is built into the database. Harmful image analysis system characterized in that the analysis of the harmful by determining whether it corresponds to the hash value of the image.
Modified contents extracting hash and DB based blocking technique and ROI feature points to determine whether the image is harmful because the hash value of the image is included in the database constructed as the hash value of at least one harmful image. An analysis server for analyzing the harmfulness of the image by using a combination of a blocking method; Wow

The hash value of the image and the feature points of the region of interest (ROI) are calculated and transmitted to the analysis server according to a request for reproducing the image, and the analysis of the harmfulness of the requested image is requested. Hazardous image analysis system comprising a user terminal for controlling the playback and playback block.
The method according to claim 5,

The analysis server,

Harmful image characterized by analyzing the harmful image by applying the hash- and DB-based blocking technique first, and analyzing the harmful image by applying the modified content-based blocking technique when the harmful image is not analyzed. Analysis system.