Underlying Feature-based Video Retrieval

Searching multimedia resource via keywords in search engines like Google,Baidu has its limitations,which can not satisfy the need of retrieving huge amount of multimedia resources.Hence,content based image/video retrieval is currently a hot research topic.This paper implements a content based video retrieval system.The system can search videos via low level features,such as color and texture.We carry out experiments on TRECVID platform to verify the performance of the system.