Image Retrieval Research Based on Broadcasting & Film
Ying Mei Han · 2007
The broadcasting film P2P content extraction middleware system based on IPv6 is a China Next Generation Internet project of National Development and Reform Commission.On the background of this project,the content based image retrieval is studied.The normal color and texture feature extraction algorithm is mainly researched.Choose the block RGB means method,advanced reference color table and grey smooth co occurrence matrix to extract the features from images.The standard Corel image library is used to set up the image feature database.In this paper,not only one single feature is used to retrieve images,but also combined-feature is used to retrieve.And the result is analyzed.Lots of experiments prove that the retrieval result gaining from combined-feature is more efficient.The recall and precision is improved.So the combinedfeature is used to retrieve the broadcasting film images.