KR101081308B1 - Method for distinguishing video copyright infringement and the apparatus using the same - Google Patents

Abstract

The present invention relates to a device for determining copyright violation by comparing a video file to be shared with a copyright file stored in a server and a method of determining using the same. To this end, the present invention provides a server for storing video asset data; And a client for transmitting shared video data to the server for sharing the video, wherein the client extracts a hash value of the shared video data and transmits the hash value to the server, and the server hashes the pre-stored video asset data. Provided is a video copyright violation determination device that compares a value with a hash value of the shared video data to determine whether there is a copyright violation.
Therefore, according to the present invention, whether a copyright violation can be reliably determined by comparing a video work and a shared video, and whether a copyright violation can be determined even when the shared video is modified to avoid copyright violation.

Description

Video copyright infringement and the apparatus using the same}

The present invention relates to a video copyright violation discrimination apparatus and a video copyright violation discrimination method using the same, and specifically, to a device for determining whether a copyright violation is made by comparing a video file to be shared with a copyright file stored on a server and a method for determining the same. It is about.

The traditional way of watching a movie or drama is to go to a movie theater or watch it on TV at home, or recently, due to the development of information and communication, various types of video files can be used to watch a movie without going directly to the movie theater. You can watch TV programs.

These video assets can be viewed and downloaded for a fraction of the time on legitimate sites, or paid and viewed at a flat rate.

However, due to the intangible nature of the software, it is often downloaded and used through illegal channels without paying legal fees. In recent years, movie files such as 'Haeundae' and 'Bat', which have been successful in Korea, have been illegally distributed and became a social problem.

Such copyright violations are particularly serious in the form of software, but there is also a problem of entrusting this reality to the user's conscience or change of consciousness.

In order to solve such a problem, a method of embedding a digital watermark in a moving image file has been studied. Digital watermark is a technology for efficiently protecting the copyright of the author by inserting information that can prove the author or the signature of the author of the digital media.

However, this requires a separate watermark insertion process and prevents illegal copying itself and has a limit in preventing illegal use through file sharing.

Accordingly, there is an urgent need for a powerful copyright protection device to reliably prevent infringement of copyright through video file sharing.

In order to solve the above problems, an object of the present invention is to provide a method for preventing copyright infringement by sharing a video asset and a video copyright violation discrimination apparatus using the same.

In order to achieve the above object, the present invention provides a server for storing video asset data; And a client for transmitting shared video data to the server for sharing the video, wherein the client extracts a hash value of the shared video data and transmits the hash value to the server, and the server hashes the pre-stored video asset data. Provided is a video copyright violation determination device that compares a value with a hash value of the shared video data to determine whether there is a copyright violation.

In addition, a server that stores the video asset data; And a client transmitting the shared video data to the server for sharing the video. (A) The client may include a reference frame image of the shared video data, a number Nc of the reference frame, Transmitting the frames per second (Rc) and the total reproduction time (Tc) to the server; (b) the server calculating Nc / Rc and storing the calculated time Fc at which the reference frame is played; (c) storing the result of multiplying the Fc by the number of frames per second Rs of the video asset data as the number of asset frames Ns of the video asset data; and (d) comparing the copyrighted frame image corresponding to the Ns with the reference frame image to determine whether a copyright violation occurs.

On the other hand, more specifically a server that stores the video asset data; And a client transmitting the shared video data to the server for sharing the video. (A) The client may include a reference frame image of the shared video data, a number Nc of the reference frame, Transmitting the frames per second (Rc) and the total reproduction time (Tc) to the server; (b) the server calculating Nc / Rc and storing the calculated time Fc at which the reference frame is played; (c) storing the result of multiplying the Fc by the number of frames per second Rs of the video asset data as the number of asset frames Ns of the video asset data; calculating a difference ΔT between the total reproduction time Ts of the video asset data and the total reproduction time Tc of the shared video data; (e) determining the number of comparison frames Cs of the video asset data to be compared with the reference frame image of the shared video data by moving in the Ns based on the magnitude and sign of ΔT; (f) it provides a video copyright violation determination method for comparing the reference frame image and the reference frame image corresponding to the Cs to determine whether the copyright violation. Here, it is preferable that (DELTA) T = Ts-Tc and Cs = Ns + ((DELTA) T * Rs).

Also, in the step (f), it is preferable to compare all the frames between the Cs and the Ns and the reference frame image.

According to the present invention, it is possible to reliably determine whether a copyright violation occurs by comparing a video work and a shared video.

In addition, it is possible to determine whether a copyright violation occurs even if the shared video is modified to avoid copyright violations.

1 is a block diagram of a server and a client showing an embodiment according to the present invention;
2 is a block diagram showing a video asset storage structure;
3 is a flowchart illustrating a sharing attempt process at a client;
4 is a flowchart illustrating a video copyright violation determination process in a server;
5 is an exemplary diagram illustrating a frame of shared video data;
6 to 11 are explanatory diagrams for explaining a video copyright violation determination process according to the present invention.

The configuration and operation of the embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Referring to FIG. 1, the system can be applied to various network construction methods. However, in the present embodiment, the system is implemented in a central server construction method of a star type.

The star type includes a server 100 and a client 200 connected radially from the server 100.

The server 100 stores a plurality of video asset files. When the client 200 attempts to share a video file, the server 100 determines whether a copyright infringement is performed in comparison with the stored video asset.

The plurality of video asset files may be stored in the server 100 in various ways. However, in the present embodiment, the plurality of video asset files are implemented in a B-Tree method, which is a data storage method formed in a tree form in which each node has children.

Each node stores one image information. Specifically, it contains the hash value and title of the video asset file, the frames per second, the total number of frames, the video playback time, the absolute path where the video is stored, and the pointer to the video asset file. . When the server program is executed, the index file representing the absolute paths of the video works stored in the server at the designated path is read, and the tree is searched to find the video works in the absolute path. Thereafter, the shared video information received from the client 200 is compared with each node to determine whether copyright infringement occurs.

The server 200 records log information in case of inaccurate determination of copyright infringement. The log information is divided into a true matching log and a false matching log. The log information is stored, for example, in a folder such as 'FriDec041015522009', which includes the order of day, time, date, and year.

The true matching log stores a log text file in which each detection process and a successful detection step are recorded, a frame file transmitted from the client 200, a frame file matched with a video work stored in the server 100, and the like. . The true matching log is a log for reviewing a detection result when a match between the video work and the shared video is incorrect.

The fall log stores each detection process and two reference frames to be described later received from the client 200 at a designated location. The false log is a log for reviewing when a detected video is infringing even though it is not detected.

Referring to FIG. 2, the video asset files stored in the server 100 are stored in a specific location.

The top root 110 stores the entire data, and as a lower root, the movie work file and the log file are stored in the Movie 130 and the Log 140, respectively. Also, the ContentIndex file is an index file that is read when the server program is first executed, and stores the absolute path of the video work and the hash value of the video work.

The Movie 130 stores a video file and information about the video file, that is, the number of frames per second, the total number of frames, the total video time, and the like for each video asset.

In the log 140, two subroots of true matching and false matching exist, and a frame matched with a process when the comparison detection of the video succeeds is stored in the true matching route, and the video matching is reversed in the false matching route. The process when the detection fails and the reference frame received from the client are stored.

Next, an operation process in the client 200 will be described with reference to FIG. 3.

When the client 200 attempts to share a video file

(S10) First, a hash value of the video file is generated (S20).

The hash value is an output value by a hash function, and the hash function outputs data as a string having a relatively small size. The hash function knows the output value but cannot infer the input value with the output value.

In this embodiment, MD5, a kind of hash function, is used. MD5 is a function that outputs 128 bits and can only check the output value, not the original input value.

After extracting the hash value, the client 200 extracts a reference frame of the shared video file (S30). The reference frame refers to a frame that is directly compared with a frame of a video work stored in the server 100. In the present embodiment, two reference frames are extracted. In this case, the reference frame extracts a frame near the first quarter point and a frame near the third quarter point.

After generating the hash value and extracting the reference frame, the hash value, information of the shared video file, and the reference frame number are transmitted to the server 100.

Next, referring to FIG. 4, a method of determining whether a video copyrighted work is infringed on the server 100 will be described.

The server 100 first compares the hash value of the shared video file received from the client 200 with the hash values of the pre-stored video works to determine whether they match. If the hash values match, the shared video file is determined to be in violation of copyright and the program ends. On the other hand, if the result of comparing the hash value of all video works does not match, proceed to the next inspection method.

If you modify the shared video file at all, the hash value changes completely. In other words, when the sharing attempter modifies the shared video file, the copyright violation cannot be discriminated by the above method. Therefore, the following inspection method directly compares the frame of the movie frame with the reference frame.

A second inspection method using a reference frame will be described with reference to FIGS. 5 to 7 as follows.

The server 100 receives a reference frame number Nc, frames per second Rc, and a total playback time Tc of the shared video data from the client 200. Then, Nc / Rc is calculated and stored as the time Fc at which the reference frame transmitted from the client is played. If the number of frames per second of the video work to be stored in the server 100 is Rs and the total playing time is Ts, the server 100 may refer to the frame number Ns of the video work to be directly compared with the reference frame. Obtained using Fc. In other words, Ns = Fc * Rs is calculated. Through this process, the reference frame of the shared video data is compared with the calculated frame of the video work.

Although there are various methods of comparing the frame with the frame, in the present embodiment, the OpenCV Library (Open Computer Vision) was used. OpenCV Library is a library focused on real-time image processing. Provides various functions related to video or image processing.

Referring to FIG. 5, the total playing time of the shared video is 10 seconds and the number of frames per second is one. In addition, the shared video includes a total of 10 frames, ranging from 'ga' frame to 'cha' frame. Among them, two reference frames are arbitrarily extracted. In the present embodiment, the 'multi' frame and the 'four' frame are extracted as the reference frame.

In this case, if the 'e' frame is removed, the total playing time is a modified video 310 as shown in FIG. However, even if this is modified, the 'multi' frame of the shared video matches the 'multi' frame corresponding to 3 seconds of the video work stored in the server 100, so that the server 100 makes a copyright violation decision.

However, when 'ga' and 'me' frames of the shared video are removed as shown in FIG. 7, the reference frame becomes a 'ma' frame and thus does not match the 'da' frame of the video work. In this case, the server 100 proceeds to the next third inspection method.

In the third inspection method, the steps for obtaining the time Fc at which the reference frame is played and the number of asset frames Ns of the video asset data are the same as in the second inspection method. Furthermore, the method calculates the difference [Delta] T between the total reproduction time Ts of the video asset data and the total reproduction time Tc of the shared video data. That is, ΔT = Ts-Tc is calculated. The reason for performing such a calculation is to assume that a deformation occurs in the total playback time by applying a deformation to the shared video.

Thereafter, the comparison frame number Cs shifted from the asset frame number Ns by the number of frames corresponding to ΔT is obtained. In other words, it is calculated as Cs = Ns + (ΔT * Rs).

Next, all frames existing between the frame corresponding to Cs and the frame corresponding to Ns are sequentially compared with the reference frame to determine whether copyright violations occur.

Referring to FIG. 8, the reference frame is a 'ma' frame and the number of asset frames Ns is three. In addition, since ΔT = 10-8 = + 2 seconds and the number of comparison frames in the video asset data 150 is 5, the corresponding comparison frame becomes a 'ma' frame. Accordingly, the server 100 compares the 'ma' frame of the video work data 150 with the 'ma' frame, which is a reference frame of the shared video data, and matches the same, thereby determining that the copyright is in violation of copyright.

9 illustrates a case where a 'bak' frame is inserted in the first half of the shared video data. At this time, since ΔT = 10-11 = -1 second and the number of comparison frames in the video asset data 150 becomes 2, the corresponding comparison frame becomes a “me” frame. Therefore, the server 100 compares the 'I' frame of the video work data 150 with the 'I' frame, which is a reference frame of the shared video data, and matches the frame.

FIG. 10 illustrates a case in which the 'b' frame is removed from the first half of the shared video data and the 't' frame is removed from the middle. At this time, since ΔT = 10-8 = + 2 seconds and the number of comparison frames in the video asset data 150 becomes 5, the corresponding comparison frame becomes a 'ma' frame. Therefore, the server 100 sequentially compares the 'ma' frame and the 'multi' frame corresponding to Ns. That is, when comparing the 'la' frame next to the 'ma' frame, it is determined as a copyright violation because it matches the 'la' frame, which is a reference frame.

11 illustrates a case in which the 'ga' and 'b' frames are removed from the first half of the shared video data and the 'bak' frame is inserted in the middle. At this time, ΔT = 10-9 = + 1 second.

For comparison with the first reference frame 'ma' frame, Ns becomes 3 as in the previous case, and the corresponding frame is 'multi' frame. In addition, the comparison frame shifted by + T, which is ΔT, becomes a 'la' frame. The server 100 compares the 'la' frame and the 'da' frame with the 'frame' which is the reference frame and does not match, and thus compares the next frame.

That is, Ns becomes 7 and the corresponding frame becomes the 'four' frame for comparison with the second reference frame 'a' frame. In addition, the comparison frame shifted by +1 second, which is ΔT, becomes 'A' frame. Accordingly, the server 100 determines that the copyright is in violation of the 'A' frame as the comparison frame Cs and the 'A' frame as the reference frame.

According to the present invention, it is possible to determine whether copyright infringement of a video work is most accurately determined through direct comparison of each frame unit, and it is possible to quickly determine whether copyright infringement of a video work is provided by providing a quick inspection method. have.

100: server 150: video asset data
200: Client 300: Shared video data

Claims (5)

delete A server storing moving picture asset data and frames per second Rs of the moving picture asset data; And
In the video copyright violation determination system including a client for transmitting the shared video data to the server for sharing the video,
(a) the client transmitting the reference frame image of the shared video data, the number Nc of the reference frame, the number of frames per second Rc, and the total reproduction time Tc to the server;
(b) the server calculating Nc / Rc and storing the calculated time Fc at which the reference frame is played;
(c) storing the result of multiplying the Fc by the number of frames per second Rs of the video asset data to be stored in the server as the number of asset frames Ns of the video asset data;
and (d) comparing the copyrighted frame image corresponding to the Ns with the reference frame image to determine whether a copyright violation occurs. A server storing moving picture asset data, frames per second Rs and total playing time Ts of the moving picture asset data; And
In the video copyright violation determination system including a client for transmitting the shared video data to the server for sharing the video,
(a) the client transmitting the reference frame image of the shared video data, the number Nc of the reference frame, the number of frames per second Rc, and the total reproduction time Tc to the server;
(b) the server calculating Nc / Rc and storing the calculated time Fc at which the reference frame is played;
(c) storing the result of multiplying the Fc by the number of frames per second Rs of the video asset data to be stored in the server as the number of asset frames Ns of the video asset data;
(d) calculating a difference ΔT between the total reproduction time Ts of the moving image data to be compared with the total reproduction time Tc of the shared video data;
(e) determining the number of comparison frames Cs of the video asset data to be compared with the reference frame image of the shared video data by moving in the Ns based on the magnitude and sign of ΔT;
and (f) comparing the comparison frame image corresponding to the Cs with the reference frame image to determine whether a copyright violation occurs. The method of claim 3,
ΔT = Ts-Tc,
Video copyright violation determination method, characterized in that Cs = Ns + (ΔT * Rs). The method of claim 3,
In the step (f), all the frames between the Cs and the Ns and the reference frame image, characterized in that for comparing the reference frame image.

Images