Enhanced Image Retrieval System Using Relavance Feedback

Abstract

In this paper a new approach for image retrieval is analyzed using image content by means of wavelet decomposition. Existing content-based image retrieval (CBIR) systems are already in use but they are useful for some particular domains. This new technique provides a efficient approach with the image content using wave let decomposition. Pyramid wavelet transform isused for image
decomposition and Daubechies family of filters is employed for noise removal and filtering operations. It is used for finding the similarity measures between the query images in the database. This will be used for sub-image matching. Calculating the
low frequency band for the image comprise many advantages to reduce memory space, because all other higher frequency bands are eliminated simulated to show the performance analysis of the new approach.