A New Semantic Approach for CBIR Based on Beta Wavelet Network Modeling Shape Refined by Texture and Color Features

摘要

Nowadays, large collections of digital images are being created. Many of these collections are the product of digitizing existing collections of analogue photographs, diagrams, drawings, paintings, and prints. Content-Based Image retrieval is a solution for information management. Image retrieval combining low level perception (color, texture and shape) and high level one is an emerging wide area of research scope. In this paper, we presented a new semantic approach based on extraction of shape refined with texture and color features extraction, using 2D Beta Wavelet Network (2D BWN) modeling. The shape descriptor is based on Best Detail Coefficients (BDC), the texture descriptor is based on Best Approximation Coefficients (BAC) and the one for color is calculated on the approximated image by applying the first two moments. Experimental results for Wang database showed the effectiveness of the proposed method.