期刊名称：International Journal of Electronics, Communication and Soft Computing Science and Engineering
印刷版ISSN：2277-9477
出版年度：2015
卷号：4
期号：Special 2
出版社：IJECSCSE
摘要：Content-based image retrieval (CBIR) is a widely used technique for retrieving images from large database. But there are many issues with this method regarding its performance. So there is scope to improve the processing. In recent years, different techniques have been implemented to improve the performance of CBIR. The proposed CBIR system uses kernel mean shift clustering, which is density based clustering. It is a nonparametric method and uses pairwise constraints to guide the clustering process. Since, Clustering is the method of an unsupervised classification but it is observed that whenever a small amount of supervision is provided to clustering, it increases the clustering performance significantly. The existing methods, such as k-mean clustering, are sensitive to initialization. Performance of the system will improve in terms of precision and recall. The performance of the proposed system is compared with the existing method
关键词：Semi-supervised kernel clustering; mean shift clustering; ; Content based image retrieval; low rank kernel learning
