基于语义的视频检索关键技术综述

随着大量视频的出现,视频内容检索是当今多媒体应用的一个重要研究方向。现有的视频检索技术多是基于低层特征,这些低层特征与高层语义概念相差较多,严重影响了视频内容检索系统的实用性。由于低层特征和高层语义概念间的语义鸿沟,如何从视频内容中提取人类思维中的语义概念,正成为目前视频内容检索中最具有挑战性的研究内容。文中介绍了语义视频检索出现的背景和国内外最新研究动态,分析了现有方法的优缺点,对现有的关键技术进行综述。

A Survey on Semantic-based Video Retrieval Key Techniques

With the emergence of much video,video content retrieval becomes an active research direction in the multimedia applications.Most of the existing video retrieval technologies are based on low-level features.These features are quite different from the semantic concepts.It seriously influences the practicality of the video content retrieval system.The gap between low-level features and high semantics is difficult to narrow,so how to extract semantic concepts in the human thought from video is becoming a most challenging research of the video content retrieval.This paper introduces the background of semantic video retrieval and the latest and dynamic research at home and abroad,analyzes the advantages and disadvantages of the existing methods and summarizes the existing key technologies.