A Novel Method for Content Based Image Retrieval Using the Approximation of Statistical Features, Morphological Features and BPN Network

Retrieving images from large and varied collections using the image texture is a challenging and important problem nowadays. This paper presents a new image representation based on the statistical properties, morphological features and fuzzy cluster features of the image. The distance measure of input query image and the database image is learned and determined through a back propagation network. The proposed method overcomes almost all the basic problems such as transition, scaling and rotation. Experimental results show that the proposed method is quite effective, both in its performance and retrieval rate.