A new system for image retrieval using beta wavelet network for descriptors extraction and fuzzy decision support

Image retrieval has been popular for several years. There are different system designs for content-based image retrieval (CBIR). So, it is very important to find effective and efficient feature extraction techniques. This paper proposes a new local approach for CBIR system which combines mechanisms including content-based image, as well as fuzzy systems. First, we exploit Beta Wavelet Network Analysis (BWNA) to extract three descriptors which are shape, texture and color. Then, we propose a fuzzy decision support system, with three inputs, for descriptors fusion and better making final decision. The experimental results show the robustness and the efficiency of the proposed system for CBIR.