KR20160116585A - Method and apparatus for blocking harmful area of moving poctures - Google Patents

Method and apparatus for blocking harmful area of moving poctures Download PDF

Info

Publication number
KR20160116585A
KR20160116585A KR1020150044495A KR20150044495A KR20160116585A KR 20160116585 A KR20160116585 A KR 20160116585A KR 1020150044495 A KR1020150044495 A KR 1020150044495A KR 20150044495 A KR20150044495 A KR 20150044495A KR 20160116585 A KR20160116585 A KR 20160116585A
Authority
KR
South Korea
Prior art keywords
face
harmful
frame
person
unit
Prior art date
Application number
KR1020150044495A
Other languages
Korean (ko)
Inventor
최진우
Original Assignee
한국전자통신연구원
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 한국전자통신연구원 filed Critical 한국전자통신연구원
Priority to KR1020150044495A priority Critical patent/KR20160116585A/en
Publication of KR20160116585A publication Critical patent/KR20160116585A/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/454—Content or additional data filtering, e.g. blocking advertisements
    • H04N21/4545—Input to filtering algorithms, e.g. filtering a region of the image
    • G06T5/001—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0042—

Landscapes

  • Engineering & Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Image Analysis (AREA)

Abstract

An apparatus for blocking the harmful area of a video determines whether the current frame of the video is a harmful frame, masks a detected area among a face and at least parts of a person except for the face after detecting the face and at least parts of the person except for the face for the harmful frame, and traces the face and at least parts of the person by using information about the face and at least parts of the person detected in a previous frame for a subsequent frame of the current frame sequentially as an initial value so as to mask the traced area.

Description

METHOD AND APPARATUS FOR BLOCKING HARMFUL AREA OF MOVING POCURES [0001]

More particularly, the present invention relates to a method and an apparatus for analyzing harmfulness of moving picture contents and blocking only a harmful area for a frame determined to be harmful at the time of moving picture reproduction in case of harmful contents will be.

While the Internet has become widely available, anyone can get information easily, but it becomes a social problem as it makes it harder to access pornographic websites, obscene video file distribution, obscene personal broadcasting and streaming services.

In order to prevent such problems, several methods are used. These methods can be classified into blocking based on URI or network address, block using pornographic information DB, detection through text analysis, and detection through image analysis. URI or network address blocking and pornographic information DB based blocking system can block in real time and it does not block legitimate websites or information, but it has a disadvantage of slow response to newly appeared websites or pornography have. On the other hand, a system that automatically analyzes text or images to determine whether pornographic material exists can detect previously unknown web sites or pornographic material, but in the environment of limited computing power and memory such as smart terminals, There is a limit to doing. The pornography discrimination method based on image analysis can not block all harmful frames for contents distributed in one file or streaming form. If the pornography discrimination method based on image analysis is judged to be harmful, the whole content is blocked and if it is judged to be harmless, the whole is reproduced. However, since there is a limit to the accuracy of the hazard judgment, blocking / regeneration of the entire contents may cause problems when the judgment result is incorrect.

A problem to be solved by the present invention is to provide a video harmful area blocking method and apparatus capable of blocking harmful frames with high speed and high accuracy based on image analysis for streaming or moving pictures in file format.

According to an embodiment of the present invention, a moving picture harmful area blocking method of a moving picture harmful area blocking device is provided. A method for blocking a moving picture harmful area includes the steps of: determining whether a current frame of an input moving picture is a harmful frame; detecting at least a part of a face and a person excluding the face when the current frame is a harmful frame; Masking the detected region of at least a portion of the face of the current frame and using information about at least a portion of the face and the face detected in the immediately preceding frame sequentially for a subsequent frame of the current frame as an initial value, And masking the tracked region of at least a portion of the face and the person.

According to an embodiment of the present invention,

For an object determined to be harmful once, masking is performed until the object disappears by using object detection and tracking, so that a high blocking rate can be achieved. In addition, even if the object is harmlessly determined as a harmful object and masked, it is possible to reduce the deterioration of the harmful blocking accuracy due to false alarm because the object is once disappeared out of the angle of view.

Therefore, according to the embodiment of the present invention, a harmful object can be blocked with high accuracy with respect to moving image contents.

1 is a flowchart illustrating a video harmful area blocking method according to an embodiment of the present invention.
FIG. 2 is a block diagram of a video harmful area blocking apparatus according to an embodiment of the present invention. Referring to FIG.
FIG. 3 is a schematic view of a moving picture harmful area blocking apparatus according to another embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification and claims, when a section is referred to as "including " an element, it is understood that it does not exclude other elements, but may include other elements, unless specifically stated otherwise.

Now, a video harmful area blocking method and apparatus according to an embodiment of the present invention will be described in detail with reference to the drawings.

1 is a flowchart illustrating a video harmful area blocking method according to an embodiment of the present invention.

Referring to FIG. 1, the moving picture harmful area blocking apparatus 100 performs the following procedure for each frame when a moving picture is input.

First, the noxious area blocking apparatus 100 of the moving image checks if the Nth frame (N = 1) is inputted (S102) and whether there is a person or face being tracked (S104). The presence of the person or face being tracked means that the human or face region has been masked because it has already been judged as a harmful image in a previous frame. In addition, the existence of a person or face being tracked means that the position of a person or a face in the previous frame is estimated and stored in the memory.

The video harmful area blocking device 100 tracks where the position of the person or the face has moved in the current frame and updates the position and size of a person or a face if a person or a face being tracked exists. The video harmful area blocking apparatus 100 regards the updated person or the face as a person who currently performs harmful actions to the face and also maintains the masking of the person or the face area.

If there is no person or face being tracked, the video harmful area blocking apparatus 100 checks whether the input frame is harmful or not using the image content-based harmfulness determination engine (S106).

If the input frame is not a harmful frame (S108), the video harmful area blocking apparatus 100 increments N by 1 (S110) and performs step S104 for the next frame.

If the inputted Nth frame is a harmful frame (S108), the video harmful area blocking apparatus 100 performs human detection and face detection on the Nth frame (S112). "Face detection" means estimating the position and size of a face including a person's eyes, nose, and mouth. "Detecting a person" means estimating the location and size of a person's body or part of a person, except the face.

When the person or the face is detected (S114), the video harmful area blocking apparatus 100 stores the detected position and size information in the memory.

In addition, the moving picture harmful area blocking device 100 masks the detection area (S116).

Next, the moving picture harmful area blocking apparatus 100 increments N by 1 (S118), and updates position and size information about a person or a face by tracking a person or a face in a next frame (S120). At this time, the moving picture harmful area blocking device 100 may update the position and size information of a person or a face by tracking a person or a face in a next frame with the position and size information detected in the immediately preceding frame as initial values.

When people or faces are traced, people sometimes disappear outside the angle of view. The moving picture harmful area shielding apparatus 100 confirms whether or not a person or a face has disappeared out of the angle of view (S122).

If the person or face does not disappear from the angle of view, the moving picture harmful area blocking device 100 masks the person or the face area tracked in the frame (S124).

On the other hand, when the person or face disappears outside the angle of view, the video harmful area blocking apparatus 100 terminates the process for the frame and performs step S104 for the frame.

The video harmful area blocking apparatus 100 repeatedly performs the above-described process for the next frame when a person or a face disappears outside the angle of view. If it is determined that a frame is harmful after repeating this process, a person or a face existing in the frame is detected and tracked to mask the corresponding area.

When a person or a face disappears outside the angle of view, it can be seen that no human exists in the frame due to camera movement or editing of the image. When a harmful image is a harmful act or a scene with a scene, a non-human image can be seen as a harmless image. Therefore, the moving picture harmful area blocking apparatus 100 according to the embodiment of the present invention stops blocking when a harmful action or a scene is cut off in the middle, and automatically stops harmful objects only when a harmful action or a scene comes later And blocks it.

On the other hand, if none of the person or the face is detected (S114), the video harmful area blocking device 100 increments N by 1 (S110) and performs step S104.

In this way, when the moving picture is input, the moving picture harmful area blocking apparatus 100 determines whether or not the corresponding frame is a harmful frame by sequentially using the video content analysis based harmfulness determination engine from the first frame. In the case of a harmful frame, the moving picture harmful area blocking apparatus 100 masks an area corresponding to a person or a face using human and face detection and tracking, and maintains masking by performing only human and face tracking in the next frame. If the frame is not a harmful frame, the video harmful area blocking apparatus 100 goes to the next frame to perform the hazard judgment.

FIG. 2 is a block diagram of a video harmful area blocking apparatus according to an embodiment of the present invention. Referring to FIG.

2, the moving picture harmful area blocking apparatus 100 includes a harmfulness determining unit 110, a human detecting unit 120, a face detecting unit 130, a human tracking unit 140, a face tracking unit 150, (160) and a memory unit (170).

When the moving picture frames are input, the harmfulness determining unit 110 stores the moving picture input frame in the memory unit 170 and determines whether the input frame of the moving picture is harmful. To this end, the hazard judgment unit 110 includes a harmfulness judgment engine (not shown). The harmfulness determination engine of the harmfulness determination unit 110 analyzes the content-based feature of the object included in the current frame to determine whether the current frame is a harmful frame or not. Content-based features may include color, shape, and texture of the image.

The human detection unit 120 performs human detection on the harmful frame. In order to achieve a high detection rate, a feature extraction method using a deformable part model (DPM) and a human / bean based support vector machine (SVM) based on a support vector machine A person classification method can be used. When a person is detected from a harmful frame, the person detecting unit 120 transmits the position and size information of the detected person to the person tracking unit 140. [ In addition, the human detection unit 120 stores the position and size information of the detected person in the memory unit 170 together with information on the corresponding frame.

The face detecting unit 130 performs face detection on a harmful frame. As a face detection method, a commonly used face detection algorithm can be used. When a face is detected from a harmful frame, the face detection unit 130 transmits the position and size information of the detected face to the human tracking unit 140. The face detector 130 stores the position and size information of the detected face in the memory unit 170 together with the information about the corresponding frame.

The person tracking unit 140 sequentially performs face tracking on the subsequent frame based on the detected person's position and size information, and displays the position and size information of the tracked person in the memory unit 170 together with information on the corresponding frame, .

The face tracking unit 150 sequentially performs face tracking on the subsequent frame based on the detected face position and size information, and displays the position and size information of the tracked face together with information on the corresponding frame in the memory unit 170. [ .

The human tracking unit 140 and the face tracking unit 150 can use an algorithm widely used in human or face tracking algorithms and can use a particle filter based method for high tracking accuracy.

In addition, the person tracking unit 140 and the face tracking unit 150 check whether the person and the face being tracked have gone out of the angle of view, and if not, the position and size information of the person and the face are displayed on the masking unit 160 . On the other hand, the human tracking unit 140 and the face tracking unit 150 terminate the person and face tracking when the person being tracked and the face disappear outside the angle of view, and deliver the end of tracking of the current frame to the masking unit 160 .

That is, the person tracking unit 140 and the face tracking unit 150 sequentially perform the person and face tracking for the subsequent frame until the person and the face being tracked disappear out of the angle of view.

The masking unit 160 masks an area corresponding to the position and size information of a person or a face. The masking unit 160 also informs the hazard judgment unit 110 of the end of tracking of the current frame when the masking unit 160 receives the tracking end from the human tracking unit 140 and the face tracking unit 150.

Then, the harmfulness judging unit 110 judges whether or not it is a harmful frame for the next frame of the tracked track.

At least some functions of the video harmful area blocking method and apparatus according to the embodiments of the present invention described above can be implemented by hardware or software combined with hardware. Hereinafter, an embodiment in which a video harmful area blocking method and apparatus are combined with a computer system will be described in detail with reference to FIG.

FIG. 3 is a schematic view of a moving picture harmful area blocking apparatus according to another embodiment of the present invention. The harmfulness determining unit 110, the human detecting unit 120, the face detecting unit 130, A human tracking unit 140, a face tracking unit 150, a masking unit 160, and a memory unit 170. As shown in FIG.

3, the video harmful area blocking apparatus 300 includes a processor 310, a memory 320, at least one storage 330, an input / output (I / O) interface 340, Interface 350. FIG.

The processor 310 may be implemented as a central processing unit (CPU) or other chipset, a microprocessor, etc., and the memory 320 may be a dynamic random access memory (DRAM), a Rambus DRAM DRAM, RDRAM), synchronous DRAM (synchronous DRAM, SDRAM), static RAM (SRAM), and the like. The storage device 330 may be a hard disk, a compact disk read only memory (CD-ROM), a compact disk rewritable (CD-RW), a digital video disk ROM, , An optical disk such as a blu-ray disk, a flash memory, various types of RAM, or a permanent or volatile storage device. The I / O interface 340 also allows the processor 310 and / or the memory 320 to access the storage device 330 and the network interface 350 may be coupled to the processor 310 and / To access the network.

In this case, the processor 310 may perform at least one of the functions of the harmfulness determining unit 110, the human detecting unit 120, the face detecting unit 130, the human tracking unit 140, the face tracking unit 150, and the masking unit 160 A program command for implementing some functions may be loaded into the memory 320 to execute a program command so that the operations described with reference to FIGS. 1 and 2 are performed. The memory 320 or the storage device 330 may also store the memory unit 170. The program command may be stored in the storage device 330 or may be stored in another system connected to the network.

The processor 310, the memory 320, the storage 330, the I / O interface 340, and the network interface 350 shown in FIG. 3 may be implemented in one computer or distributed in a plurality of computers .

The embodiments of the present invention are not limited to the above-described apparatuses and / or methods, but may be implemented through a program for realizing functions corresponding to the configuration of the embodiment of the present invention or a recording medium on which the program is recorded, Such an embodiment can be readily implemented by those skilled in the art from the description of the embodiments described above.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, It belongs to the scope of right.

Claims (1)

A video harmful area blocking method of a video harmful area screening device,
Determining whether a current frame of the input moving image is a harmful frame,
Detecting at least a portion of a face and a person excluding the face when the current frame is a harmful frame,
Masking the detected region of the face and at least a portion of the person,
Tracking at least a portion of the face and the person using the detected face in the immediately preceding frame and information about at least a portion of the person as an initial value for a subsequent frame of the current frame,
Masking the tracked area of the face and at least a portion of the person
The method comprising the steps of:
KR1020150044495A 2015-03-30 2015-03-30 Method and apparatus for blocking harmful area of moving poctures KR20160116585A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020150044495A KR20160116585A (en) 2015-03-30 2015-03-30 Method and apparatus for blocking harmful area of moving poctures

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020150044495A KR20160116585A (en) 2015-03-30 2015-03-30 Method and apparatus for blocking harmful area of moving poctures

Publications (1)

Publication Number Publication Date
KR20160116585A true KR20160116585A (en) 2016-10-10

Family

ID=57146066

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020150044495A KR20160116585A (en) 2015-03-30 2015-03-30 Method and apparatus for blocking harmful area of moving poctures

Country Status (1)

Country Link
KR (1) KR20160116585A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20180071156A (en) * 2016-12-19 2018-06-27 삼성전자주식회사 Method and apparatus for filtering video
KR102169073B1 (en) 2019-09-30 2020-10-22 광운대학교 산학협력단 Real-time Streaming Blocking System and Method Based on the Results of the Video Obscenity Analysis
KR102387441B1 (en) 2020-11-11 2022-04-14 광운대학교 산학협력단 Streaming Control System based on Video Stream Characteristics for Real-time Response to Harmful Video Streaming
CN115086738A (en) * 2022-06-07 2022-09-20 广州博冠信息科技有限公司 Information adding method and device, computer equipment and storage medium
US11470385B2 (en) 2016-12-19 2022-10-11 Samsung Electronics Co., Ltd. Method and apparatus for filtering video

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20180071156A (en) * 2016-12-19 2018-06-27 삼성전자주식회사 Method and apparatus for filtering video
US11470385B2 (en) 2016-12-19 2022-10-11 Samsung Electronics Co., Ltd. Method and apparatus for filtering video
KR102169073B1 (en) 2019-09-30 2020-10-22 광운대학교 산학협력단 Real-time Streaming Blocking System and Method Based on the Results of the Video Obscenity Analysis
KR102387441B1 (en) 2020-11-11 2022-04-14 광운대학교 산학협력단 Streaming Control System based on Video Stream Characteristics for Real-time Response to Harmful Video Streaming
CN115086738A (en) * 2022-06-07 2022-09-20 广州博冠信息科技有限公司 Information adding method and device, computer equipment and storage medium

Similar Documents

Publication Publication Date Title
US10846867B2 (en) Apparatus, method and image processing device for smoke detection in image
KR20160116585A (en) Method and apparatus for blocking harmful area of moving poctures
CN103312770B (en) Method for auditing resources of cloud platform
WO2017166494A1 (en) Method and device for detecting violent contents in video, and storage medium
US20190304112A1 (en) Methods and systems for providing selective disparity refinement
WO2021254134A1 (en) Privacy shielding processing method and apparatus, electronic device, and monitoring system
KR101704830B1 (en) Apparatus and method for detecting object in image
CN111383168A (en) Privacy-protecting camera
WO2016197657A1 (en) Photographing processing method and apparatus
CN110708568B (en) Video content mutation detection method and device
KR20160037480A (en) Method for establishing region of interest in intelligent video analytics and video analysis apparatus using the same
US10614336B2 (en) Method, system, and computer-readable recording medium for image-based object tracking
KR20110066676A (en) Apparatus and method for blocking the objectionable multimedia based on skin-color and face information
US11587202B2 (en) Method, apparatus or computer program for controlling image processing of a captured image of a scene to adapt the captured image
KR101840042B1 (en) Multi-Imaginary Fence Line Setting Method and Trespassing Sensing System
JP2011070629A (en) Advertising effect measurement system and advertising effect measurement device
KR101925799B1 (en) Computer program for preventing information spill displayed on display device and security service using the same
CN108334811B (en) Face image processing method and device
CN113095272B (en) Living body detection method, living body detection device, living body detection medium and computer program product
WO2023273132A1 (en) Behavior detection method and apparatus, computer device, storage medium, and program
CN112752110B (en) Video presentation method and device, computing device and storage medium
KR101908938B1 (en) Detection of foreground object using background model-based temporal minimum filtering and log histogram
CN113947795A (en) Mask wearing detection method, device, equipment and storage medium
JP2003143593A (en) Monitor method and system utilizing picture processing
JP2008085539A (en) Program, detection method, detector, image processing method and image processor