With the rapid development of internet and digital technology, the importance of video data in the network has become increasingly prominent, and users' demand for video retrieval has increased. In view of the limitations of traditional video retrieval methods in mining video content, this paper designs a television video retrieval system that integrates multi-modal information, uses deep learning technology to extract image and text information from video, and establishes a retrieval model for information fusion, and finally establishes a storage server for the fused information to achieve more accurate and comprehensive video retrieval.
关键词
多模态信息/电视视频/检索系统
Key words
multimodal information/television video/retrieval system
维普
万方数据