JP2002171481A - Video processing apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make a job for attaching an index to a video image efficient so as to reduce the time and the cost required for the job. SOLUTION: A video processing apparatus editing and retrieving a video image is provided with a video input section 101 that receives a structured video image, an audio recognition section 102 that recognizes audio in the video image, a keyword extract section 103 that extracts a keyword from the recognized audio, and a keyword recording section 104 that records the keyword together with the cross-reference with respect to the video image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video processing apparatus for further editing or retrieving a structured video.

[0002]

2. Description of the Related Art In recent years, digital data (hereinafter, referred to as video in this specification) including moving images and audio has been actively distributed due to improvements in recording media technology and communication technology. With the distribution of video, instead of simply displaying the video recorded on the recording media in chronological order,
There is a need for a system that can extract, view, and edit the necessary parts of the video.

[0003] Currently, there is a system for extracting, browsing, and editing a video in which the contents expressed by the video are described as an index, and the video is searched, reconstructed, and re-edited based on the index information. In this specification, a process of classifying a video into a groove to which an index is added is referred to as video structuring. Structuring is often performed by hierarchically classifying images.

[0004] There are two methods for indexing a video image: a physical description method and a linguistic description method. The physical description method is a method of describing contents according to rules such as scene switching and camera angle.
The linguistic description method is a method of describing, for example, the name of a subject in a video, the behavior of the subject, and the like in text.

[0005]

The simplest method of indexing with a linguistic description method is manual. However, the task of manually indexing a large number of videos with a linguistic description method has a disadvantage that the amount of work is large, and the time and cost required for the work are enormous. Therefore, there is a demand for a video processing device that can efficiently attach an index reflecting the content to a video.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and has as its object to provide an image processing apparatus capable of efficiently indexing a video and reducing the time and cost required for the operation. And

[0007]

[MEANS FOR SOLVING THE PROBLEMS] To solve the above-mentioned problems,
According to another aspect of the present invention, there is provided a video processing apparatus for editing and retrieving a structured video, wherein the video input unit inputs the structured video. Voice recognition means for recognizing the voice of the video input to the video input means, keyword extraction means for extracting a keyword from the voice recognized by the voice recognition means, and correspondence between the keyword and the structured video Keyword recording means for recording with the relationship;
It is characterized by having.

According to the first aspect of the present invention, a keyword can be automatically attached to a structured video based on audio included in the video. For this reason, the load on the worker can be greatly reduced and the work time can be reduced as compared with the case where the keyword is manually added (the task of attaching the keyword while watching the video).

A video processing apparatus according to a second aspect of the present invention further comprises: a title selection unit that selects a title of the structured video from one of the keywords extracted by the keyword extraction unit; Title recording means for recording the title together with the correspondence to the structured video.

According to the second aspect of the present invention, a title can be automatically given to a structured video based on a keyword.

According to a third aspect of the present invention, in the video processing device, the title selecting means selects a keyword having a high frequency of appearance in a predetermined unit from among the keywords extracted by the keyword extracting means. It is characterized by.

According to the third aspect of the present invention, more keywords can be reflected in the title of the video.

[0014] The video processing apparatus according to the invention of claim 4 further comprises candidate registration means for pre-registering a keyword candidate or a title candidate, and among the information included in the audio of the structured video. In the case where information that matches a keyword or title registered in the candidate registration means is included, candidate selection means for selecting the information as a keyword or title is provided.

According to the fourth aspect of the present invention, a keyword or a title can be selected from keyword or title candidates registered in advance.

According to a fifth aspect of the present invention, in the video processing apparatus, the registering means registers a keyword candidate or a title candidate with selection priority.

According to the fifth aspect of the present invention, it is possible to register the title candidate words in the registration means with the priority of selection.

A video processing apparatus according to a sixth aspect of the present invention is characterized in that the keyword recording means records the keyword as text data or audio data.

According to the invention described in claim 6, the keyword recording means can record the keyword as text data (words) or as voice data.

According to a seventh aspect of the present invention, in the video processing apparatus, the title recording means records the title as text data or audio data.

According to the seventh aspect of the present invention, the title recording means can record the title as text data (words) or as audio data.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments 1 to 4 of an image processing apparatus according to the present invention will be described below in detail with reference to the accompanying drawings. (Embodiment 1) FIG. 1 is a block diagram for explaining a configuration of a video processing apparatus according to Embodiment 1 of the present invention. The illustrated configuration is a video processing device that edits and searches a structured video. Then, a video input unit 101 for inputting the structured video, a voice recognition unit 102 for recognizing the voice of the input video, a keyword extraction unit 103 for extracting a keyword from the recognized voice, and a structured keyword And a keyword recording unit 104 that records the information together with the corresponding relationship to the video.

The structured video is input from the video input unit 101 to the video processing device of the present embodiment. The structured video is composed of a plurality of segments. The voice recognition unit 102 recognizes voice included in the video using a voice recognition technology. The keyword extracting unit 103 extracts a keyword for each segment from the recognized speech.

For example, as will be described later, keyword candidates are recorded in advance in the video processing apparatus, and if there is a keyword that matches a keyword candidate recorded in a phrase included in the sound, as described later, A phrase that matches this candidate may be extracted as a keyword. Also, for example,
A phrase that is recognized many times in voice may be used as a keyword.

According to the first embodiment, it is possible to automatically attach a keyword to a segment constituting a structured video based on audio included in the video. According to the first embodiment, it is possible to greatly reduce the load on the worker and to shorten the work time as compared with the case where the keyword is manually added (work in which the worker attaches the keyword while watching the video). it can.

(Embodiment 2) Next, a video processing apparatus according to Embodiment 2 will be described. FIG. 2 is a block diagram for explaining the video processing device according to the second embodiment. In addition, among the configurations illustrated in FIG. 2, the same configurations as the configurations illustrated in FIG. 1 are denoted by the same reference numerals, and a description thereof is partially omitted.

The video processing apparatus according to the second embodiment further includes:
A title selection unit 201 for selecting a title of a structured video from any of the keywords extracted by the keyword extraction unit 103, and a title recording unit for recording the selected title together with the correspondence to the structured video 202.

In the above-described video processing apparatus, the voice input to the video input unit 101 together with the video and the voice recognized by the voice recognition unit 102 is subjected to keyword extraction by the keyword extraction unit 103. The extracted keywords are input to the title selection unit 201. The title selection unit 201 selects a title from the extracted keywords.

The selection of a title can be performed, for example, by selecting a keyword having a high extraction frequency among the extracted keywords. When a title is selected based on the keyword extraction frequency, as shown in FIG.
May be further provided with an appearance frequency calculation unit to calculate the keyword extraction frequency. In addition, as described later, title candidates are recorded in advance in the video processing apparatus, and if there is a title that matches the candidate recorded in the keyword, the keyword that matches the candidate is extracted as a title. Is also good.

According to the second embodiment, a title can be automatically given to a structured video based on a keyword. In the second embodiment, the burden on the operator is significantly reduced as compared with the case where the title is manually assigned (the operation in which the operator views the video and assigns the title).
Work time can also be reduced.

According to the second embodiment, the keywords of more segments can be reflected in the title of the video. For this reason, a title reflecting the contents of the segments constituting the video can be given to the video.

(Embodiment 3) Next, a video processing apparatus according to Embodiment 3 will be described. FIG. 4 is a block diagram for explaining the video processing device according to the third embodiment. In addition, among the configurations illustrated in FIG. 4, the same configurations as the configurations illustrated in FIG. 1 are denoted by the same reference numerals, and description thereof is partially omitted.

The video processing apparatus according to the third embodiment further includes:
A registration unit 403 in which keyword candidates (keyword candidate phrases) or title candidates (title candidate phrases) are registered in advance, a keyword / title collation unit 401, and a keyword / title recording unit 402 are provided. The keyword / title collation unit 401 extracts a keyword that matches the keyword phrase as a keyword when the keyword included in the structured video sound includes a keyword phrase or a phrase that matches the title candidate phrase. I do. In addition, a phrase that matches the title phrase is extracted as a title (the title may be the same as the title phrase of the extracted keywords).

In the above-described video processing apparatus, the voice input together with the video to the video input unit 101 and the voice recognized by the voice recognition unit 102 is collated by the keyword / title collation unit 401 with the keyword registered in the registration unit 403. . Then, when the voice includes a phrase that matches the keyword candidate phrase in the registration unit 403, the matching phrase is extracted as a keyword of a segment including the phrase.

In the third embodiment, a title is extracted from the extracted keywords. For this reason, when there is a keyword that matches the title candidate phrase among the extracted keywords, the keyword / title collation unit 401 sets this keyword as the title.

The keywords and titles extracted by the keyword / title collation unit 401 are output to the keyword / title recording unit 402 and recorded in the keyword / title recording unit 402. It should be noted that the video processing apparatus of the present invention can also set a title even when a word that matches a title candidate word is included in the voice recognized by the voice recognition unit 102, even when this word is not extracted as a keyword. .

According to the third embodiment, by selecting a keyword or a title from candidates for a keyword or a title registered in advance, the intention of the operator can be reflected on the keyword or the title.

In the third embodiment, the keyword candidate words and the title candidate words can be registered in the registration unit 403 with the priority of selection. By doing so, the intention of the operator can be reflected on the keyword or title.

Further, in the first to third embodiments, the keyword recording unit 104, the title recording unit 202, and the keyword / title recording unit 402 record keywords and titles as text data (phrases) and also as audio data. You can also. In this way, keywords and titles can be attached to video as audio in addition to words and phrases.

[0039]

As described above, according to the first aspect of the present invention, the load on the worker is significantly reduced as compared with the case where the keyword is manually added (the task of attaching the keyword by viewing the video by the worker). This has the effect of reducing the number of working hours. The invention described in claim 1 streamlines the work of indexing a video,
It can be said that the time and cost required for the work can be reduced.

According to the second aspect of the present invention, the burden on the operator is greatly reduced and the work time is reduced as compared with the case where the title is manually assigned (the task of giving the title while watching the video). It has the effect that it can be done.
According to the second aspect of the present invention, it can be said that the work of adding an index to a video can be made more efficient, and the time and cost required for the work can be reduced.

The third aspect of the invention has an effect that a title reflecting the keyword attached to the video can be provided to the video.

According to the fourth aspect of the present invention, the intention of the operator can be reflected on the keyword or the title.

According to the fifth aspect of the invention, the intention of the operator can be reflected on the keyword and the title.

According to a sixth aspect of the present invention, a keyword is
In addition to words and phrases, they can be attached to video as audio.

According to the seventh aspect of the present invention, a title can be attached to a video as audio in addition to a word.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a video processing device according to a first embodiment of the present invention.

FIG. 2 is a block diagram illustrating a configuration of a video processing device according to a second embodiment of the present invention.

FIG. 3 is a block diagram for explaining a configuration of another video processing device according to the second embodiment of the present invention;

FIG. 4 is a block diagram illustrating a configuration of a video processing device according to a third embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 101 video input unit 102 voice recognition unit 103 keyword extraction unit 104 keyword recording unit 201 title selection unit 202 title recording unit 401 keyword / title collation unit 402 keyword / title recording unit 403 registration unit

Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat II (Reference) G10L 15/10 G10L 3/00 551G 15/00 H04N 5/91 N H04N 5/93 5/93 Z

Claims (7)

[Claims] 1. A video processing apparatus for editing and retrieving a structured video, comprising: a video input unit for inputting the structured video; and recognizing audio of the video input to the video input unit. Voice recognition means; keyword extraction means for extracting a keyword from the voice recognized by the voice recognition means; and keyword recording means for recording the keyword together with the correspondence to the structured video. Video processing device. 2. A title selecting unit for selecting a title of the structured video from one of the keywords extracted by the keyword extracting unit, and a correspondence between the title and the structured video. The video processing apparatus according to claim 1, further comprising: a title recording unit that records the title and the title together. 3. The title selection unit according to claim 1, wherein the title selection unit selects a keyword having a high appearance frequency in a predetermined unit from among the keywords extracted by the keyword extraction unit as a title. Video processing device. 4. The apparatus according to claim 1, further comprising a candidate registering means for registering a keyword candidate or a title candidate in advance, wherein information included in the structured video sound includes a keyword or a keyword registered in the candidate registering means. The video processing apparatus according to claim 1, further comprising: a candidate selecting unit that selects the information as a keyword or a title when the information including the title is included. 5. The video processing apparatus according to claim 4, wherein said registering means registers keyword candidates or title candidates with selection priorities. 6. The video processing apparatus according to claim 1, wherein the keyword recording unit records the keyword as text data or audio data. 7. The video processing apparatus according to claim 2, wherein the title recording unit records the title as text data or audio data.