Video retrieval

With the development of computer technology and network technology, the construction of information superhighway, and the popularization and application of multimedia, all kinds of video materials are continuously produced. More and more video databases have been established, and many new service forms and means of information exchange have appeared, such as digital library, digital museum, digital TV, video on demand, distance education, distance medical treatment and so on.

In traditional database systems, the retrieval of information is generally based on numerical and character types, but in multimedia databases, unformatted information such as image, video, audio and so on is integrated. They have the characteristics of large amount of data, variable length of information, and complex structure.

Video retrieval is to find the desired video segment from the massive video data. Based on the given examples or feature descriptions, the system can automatically find the desired video clips. Depending on the submitted video content, video retrieval is generally divided into shot retrieval and clip retrieval.

At present, most of the research on video retrieval still focuses on shot retrieval. However, research on segment retrieval is just beginning. From the user’s point of view, their query to a video database is usually a video clip and rarely a single physical shot. From the perspective of information content, a video clip composed of several shots has more semantics than a single shot, and it can represent the events that users are interested in. Therefore, the results of the query are also more meaningful.

Fragment retrieval is divided into the following two types: exact retrieval and similarity retrieval. The key technologies of a complete video retrieval system mainly include key frame extraction, image feature extraction, image feature similarity measurement, query method and video clip matching.

A number of content-based video retrieval systems have been developed at home and abroad, mainly as follows:

• Query By Image Content
• Visual Seek System
• Video Q
• Tsinghua Video Find It

The purpose of content-based video analysis and retrieval is to process, analyze and understand video content by computer, and build structure and index to achieve convenient and effective access to video information. According to the content and context of the video, it retrievales the large-scale video data.

Your task

What are the applications of video retrieval in life?

Share your thoughts and ideas in the comments below.