Video retrieval using singilar value decomposition and latent semantic indexing

Abstract

Similarity matching algorithm plays an important role in Video retrieval system. Most of the video retrieval systems are designed using traditional similarity matching algorithms that are based on distance measures. As the Accuracy of retrieval system depends on the method used for detecting shots, kind of features we use for retrieval, it is also mandatory to think about the algorithms that we use for similarity matching of video clips. Here we propose a new technique latent semantic indexing technique which is based on Singular value decomposition. A key feature of LSI is its ability to establish associations between similar kind of information, It easily provides the accurate index terms that matches with the query clip index. We compare the results obtained with results of traditional Euclidean Distance method. Results are satisfactory.