Indexing Video Using Global and Local Features

Video is the multimedia information needed more on Internet. For its massive volume and linear structure, video has received more and more emphasis by researchers. The corresponding retrieval method is different from traditional key-word retrieval for text information. This paper divides the video features into two types—local feature and global feature, and systematically expatiates how to extract local and global features from video and how to establish local and global feature-based model for video retrieval.