An Efficient Content Based Image Retrieval System using the Mesh-of-Trees Architecture

With the fast development of the Internet and computer technology in recent years, the use of visual information on the World Wide Web has become more popular. There has been a trend in retrieving images on the basis of automatically-derived features such as color and shape. Conducting the image retrieval by query image has attracted plenty of attention from researchers in the field of information technology, The CBIR (Content Based Image Retrieval) technique refers to the image retrieval operation based on the key primitives. Despite the fact that the CBIR technique has been accepted and implemented in many application domains, extracting images of their contents is really an arduous job. The difficulty is determining how to organize these image features in some fashion without consuming too much time. In this paper, a novel MOT (Mesh Of Trees) based CBIR system is proposed. It makes use of dual image signatures to more efficiently provide better query results. Images are preprocessed and indexed by means of both color histogram and wavelet image signatures, and "similar" images are adjacently stored in a two-dimensional MOT. With superb traversing capabilities in rows and column trees, the architecture is very effective in accommodating dual image features. For higher dimensional image feature vectors and massive image databases, the experimental results show that the proposed architecture achieves its superior performance figures in comparisons of other existing systems.