JP2002064784A - Video information structural device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a video information structural device that integrates various methods for interpretation and structuralization of video information and selects structural rules, depending on types of the video information. SOLUTION: Video structural rules corresponding to contents of a received video signal (types of video information) are selected from among different video structural rules stored in storage means and given to a video structural means. The video structural means outputs an instruction to each video information interpretation means, according to the received video structural rules, and each video information interpretation means outputs result of interpreting the video information to the video structural means according to the instruction received from the video structural means. Then the video structural means interprets and structures the video information, based on the interpretation result of the video information received from each video information interpretation means and the video structural rules.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video information structuring apparatus for interpreting video information and automatically structuring the video information.

[0002]

2. Description of the Related Art A conventional video signal is simply one-dimensional information on a time axis of video and audio, and its reproduction method is basically only simple reproduction. However, in recent years, there has been an increasing demand for efficient search, editing, processing, and quick viewing of video information, and conventional one-dimensional video signals are insufficient for performing such operations flexibly. Therefore, a method of structuring video information and adding an index indicating its content has begun to be used.

The task of interpreting and structuring video information is
Conventionally, it has been performed manually. However, this is time consuming and costly. Therefore, much research has been conducted on automatic interpretation and structuring of video information.

For example, in Japanese Patent Application Laid-Open No. 10-257436, "Automatic Hierarchical Structure Method of Moving Image and Browsing Method Using It", a moving image is divided into shots, and shots are shot using similarity of the divided shots. Are integrated to form a scene. In this method, a moving image is divided into shots by detecting a cut point based on a correlation between frames, and the shots are integrated by using the similarity of the color histogram of a representative frame of each shot to extract a scene. In addition to the above-described method using the features of images, a video interpretation method using voice recognition, caption recognition, or the like has been studied.

[0005]

The prior art attempts to interpret and structure video information using specific characteristics of a video signal, such as, for example, similarity between pixel values and color histograms. . However, the interpretation of video information
There are many features that can be used for structuring, and it is thought that high-precision structuring can be realized by combining them well. Furthermore, for example, "news""sports"
Type of video information such as "drama" (contents of video signal)
Therefore, it is considered that the policies such as the optimal combination and parameters are different. An object of the present invention is to provide a video information structuring apparatus that integrates various methods for interpreting and structuring video information and that can select a structuring rule according to the type of video information.

[0006]

Therefore, in order to solve the above-mentioned problems, the present invention provides a storage means storing a plurality of different video structuring rules, and a plurality of video memories for interpreting video information of an input video signal. Information interpreting means, and a video structuring rule selected according to the content of the input video signal is supplied from the storage means, and the plurality of video information interpreting means are controlled based on the selected video structuring rule. Video structuring means for obtaining an interpretation result of the video information, and structuring the video information based on the interpretation result of the video information and the selected video structuring rule. A structuring device.

[0007]

FIG. 1 is a block diagram showing the concept of a video information structuring apparatus according to the present invention. According to the present invention, a video signal is input to a plurality of video information interpreting means, and a video structuring rule corresponding to the content (type of video information) of the input video signal is stored in a plurality of video information interpreting means. Select from different video structuring rules and input to the video structuring means. (The user may directly select the video structuring rules.)

[0008] The video structuring means outputs an instruction to each video information interpreting means in accordance with the input video structuring rule,
Each video information interpreting means interprets the video information in accordance with the command input from the video structuring means and outputs a result to the video structuring means. Then, the video structuring means interprets and structures the video information based on the video structuring rules and the interpretation result of the video information input from each video information interpreting means.

An embodiment of the present invention will be described below. FIG. 2 is a block diagram showing an embodiment of the present invention. The input video signal is input to the video signal analyzer 11. The video signal analysis unit 11 includes a plurality of video signal analysis units 1 to 3. Each video signal analyzing means analyzes the input video signal and outputs the result to the video information interpreting section 12. The video information interpreting section 12 is composed of a plurality of video information interpreting means 1-3. Each video information interpreting means interprets the analysis result inputted from the video signal analyzing means based on the control signal inputted from the video structuring means 13 and outputs the result to the video structuring means 13.

The video structuring rule descriptions 1 to 3 are text descriptions of video structuring rules, and a plurality of different rule descriptions corresponding to the types of video information are prepared in the storage means 15. The video structuring rule descriptions 1 to 3 are selectively input to the parsing means 14.

The video structuring rule description syntax definition is a text description defining the syntax of the video structuring rule description, and is stored in the storage unit 16. The video structuring rule description syntax definition is accompanied by video information interpreter input / output specifications 1 to 3 which describe input / output specifications of each video information interpreter of the video information interpreter 12. The video structuring rule description syntax definition is input to the syntax analysis means 14.

The parsing means 14 parses the video structuring rule description based on the video structuring rule description syntax definition, outputs video structuring rule data to the video structuring means 13, and outputs an error message. .

The video structuring means 13 includes a parsing means 14
Based on the input video structuring rule data, a control signal is output to the video information interpreting unit 12, and an interpretation result corresponding to the control signal is obtained from the video information interpreting unit 12. Then, the video structuring unit 13 structures the video information based on the interpretation result input from the video information interpreting unit 12 and the video structuring rule data, and outputs the structured data to the storage device 17 (database). .

The storage device 17 is for recording structured data input from the video structuring means 13 and a video signal. The display / reproducing means 18 displays and reproduces the video signal and the structured data recorded in the storage device 17 in various forms.

Next, each component will be described in more detail. The video signal analysis unit 11 performs various kinds of signal processing before interpreting the video information. Each of the video signal analyzing means 1 to 3 performs a different type of signal processing. The signal processing performed here is for directly operating image data and audio data of a video signal such as calculation of a color histogram, motion estimation, edge detection, and phoneme recognition. Any signal analysis means necessary to interpret the video information can be provided.

The video information interpreting section 12 uses the analysis result input from the video signal analyzing section 11 to interpret video information at a higher level. Each of the video information interpreting units 1 to 3 performs different types of information interpretation using an output from an arbitrary video signal analyzing unit. The information interpretation performed here can be directly used for video structuring such as cut point detection, shot similarity calculation, object recognition, voice recognition, and the like. Since it is considered that the processing here takes a relatively long calculation time, it is possible to input a control signal from the video structuring unit 13 and process only a necessary portion of the video information. In addition, parameters and the like can be input from the video structuring unit 13.

The video structuring rule description stored in the storage means 15 describes in text a video structuring rule including control information of the video information interpreting means and various parameters used for video structuring. A unique grammar may be used for the description, and XML (Extensible M
arkup Language) may be used. Since the video structuring rule to be used is considered to be different depending on the type of video such as “news”, “sports”, “drama”, etc., the video structuring rule description to be used is selected according to the input video information.

The video structuring rule description syntax definition stored in the storage means 16 is a text description defining the syntax of the video structuring rule description, and is accompanied by input / output specifications corresponding to each video information interpreting means. I have. For example, when XML is used for the video structuring rule description, the syntax definition is DT
D (Document Type Definition)
n) can be used.

The syntax analysis means 14 analyzes the input video structuring rule description in accordance with the video structuring rule description syntax definition, converts it into structured binary data, and outputs it to the video structuring means 13. If there is an error in the description of the video structuring rule, an error message is output, and in some cases, the video structuring operation is stopped. The user can check the error message and correct the video structuring rule description if necessary.

The video structuring means 13 is a syntactic analysis means 1
The video information is structured according to the binary data of the video structuring rule input from step 4. A control signal and a parameter are output to each video information interpreting unit in accordance with the video structuring rule, and data of an interpretation result is received. It is also possible to output data input from a certain video information interpreting means as a parameter to another video information interpreting means. The data received from each video information interpreting means is processed according to the video structuring rules to structure the video information. The structured data is obtained by dividing video information hierarchically into scenes, expressing the hierarchical structure using a tree structure or the like, and adding related information to each scene. Further, related information may be added to a certain moment of the video. The structured data of the video information is output to the storage device 17.

The storage device 17 is a digital storage medium such as a hard disk, for example, and stores video signals and structured data of video information. The correspondence between the scene and time information of the structured data and the position of the video signal is also recorded.

The display / reproducing means 18 is for displaying and reproducing the video signal and the structured data of the video information recorded in the storage device 17 effectively. For example, by displaying a representative frame of a video scene hierarchically according to structured data, a user can visually grasp the entire structure of video information. Further, it is possible to designate one of the representative frames and reproduce the corresponding video scene. Alternatively, it is also conceivable to change the playback speed according to the content of the video scene during fast-forward playback. Alternatively, a video scene can be searched using a keyword or the like as an input.

As described above, the video information structuring apparatus according to the present embodiment can select a video structuring rule according to the type of video. Therefore, an optimal video interpretation / structuring method corresponding to the type of video is used. be able to.・ By describing video structuring rules in text,
The user can easily modify the video structuring rules. By using XML for the video structuring rule description, a ready-made XML parser can be used for the syntax analysis means, so that the cost of system construction can be reduced. A structure in which the video structuring unit controls each video information interpreting unit allows various video information interpreting methods to be integrated and handled. The structure in which the video information analysis unit uses the video signal analysis unit allows a plurality of video information interpretation units to share the same video signal analysis unit. It has an excellent effect.

[0024]

As described above, the video information structuring apparatus of the present invention can select a video structuring rule according to the type of video, and therefore uses an optimal video interpretation and structuring method corresponding to the type of video. be able to. In addition, since the video structuring unit controls each video information interpreting unit, various video information interpretation methods can be integrated and handled, and more accurate video information structuring can be performed.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating the concept of the present invention.

FIG. 2 is a configuration diagram showing one embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Video signal analysis part 12 Video information interpretation part 13 Video structuring means 14 Syntax analysis means 15, 16 Storage means 17 Storage device 18 Display / reproduction means

Claims (1)

[Claims] 1. A storage means in which a plurality of different video structuring rules are stored; a plurality of video information interpretation means for interpreting video information of an input video signal; The video structuring rule selected accordingly is supplied, and the plurality of video information interpreting means are controlled based on the selected video structuring rule to obtain an interpretation result of the video information. Video structuring means for structuring the video information based on the selected video structuring rule.