Image Retrieval Using Hybrid CRSFS Discriminate Pattern Selection Descriptor in Caltech 101 Database

Image Retrieval is very one of the biggest task in the recent years. It is widely used in many real time databases to retrieve related images in various fields like medical, military, online shopping etc. This paper offers with using radon transform followed by PCA and LPP techniques for image retrieval is called as Combined Radon Space Features Set (CRSFS). Caltech 101 database image sets used in this paper. The radon transform used in FFT based slice theorem. The correct direction is select means the computation time and complexity of operation is less to achieve good retrieval rate. Here PCA is used to reduce the dimensionality in feature vectors produced by radon transform and LPP is used to find the set of basic vectors which maximizes the radio between-class scatter and within-class scatter. In order to verify our method to various data set in different condition like rotation, scaled and noisy query environments like additive, Gaussian, salt and pepper. And our proposed method was achieving better retrieval rate.