Image classification using neural network for efficient image retrieval

Traditional keyword based image retrieval systems has become inefficient for retrieval of images because of extensive digitalization of images and wide explosion of World Wide Web. To overcome such limitations Content Based Image Retrieval systems are used to retrieve similar images from large database for a given query image. There are various methods for CBIR are available some of which used Global image features such as Color, Texture and Shape. Some methods uses region level image features such as image segments. In our system we are using hybrid approach. We uses global image features based CBIR with feed forward back-propagation neural network. Neural network is used for classification of query image as per training database. At first neural network is trained about the color features of images in the database. The training is done by using back-propagation algorithm. This trained database is used for classification of the query image. According to retrieved image class further color based CBIR is used for retrieving similar images.