Texture Image Retrieval Using Greedy Method

Abstract

There is a huge amount of methods for extracting image descriptors and defining the similarity measures. In this paper, we try to improve texture image retrieval performance with post processing based on the greedy technique called Prims algorithm. In the proposed method feature database is represented using distance matrix, which is the distance between every image of the database. Due to symmetric property of a matrix, we can improve the efficiency and effectiveness of the proposed retrieval system. However for large database the size of the matrix is large. The proposed system is tested with three different image descriptors, namely combined rotated complex wavelet filters (RCWF) and dual tree complex wavelets (DT-CWT), Contourlet Transform (CT), and Discrete Wavelet Transforms (DWT).