Signature Based Approach For Image Retrieval Using Color Histogram And Wavelet Transform

Abstract

Abstract— In this article, a method is proposed for image mining based on analysis of color and texture properties of an image. Color and texture are the primitive image descriptors in content based image retrieval systems. We first resize all the database images to obtained similar space. For color based Image Retrieval, HSV color model is used to obtain color histogram of image and for texture extraction Haar wavelet transform is used. Then the proposed method evaluates different number of signatures for HSV and wavelet coefficients. Similar process is applied on the submitted query image. Finally Euclidian distance between query image and database images is calculated and the images having minimum distances are extracted from the database as results.