Exploring differential evolution algorithm for content based image retreival system

Content Based Image Retrieval (CBIR) system receives paramount importance now days. This is because of its wide applicability found in many areas including medical, science, security, Bioinformatics and entertainments. It has emerged as one of the growing field of research in engineering and the sciences. CBIR system searches large image database based on the contents of the images. The recent work carried out in this field is focused to achieve efficiency and accuracy in the image retrieval process. The major limitations encountered during review of the literature are an effective representation of image by extracting visual contents, mismatches found due to semantic gap between image representation and user's interpretation of the image and high dimensional feature vectors. This paper mainly focuses on two problems, feature extraction techniques to effectively represent the image and semantic gap. To attenuate the problems identified, the paper presents a proposal of the novel framework to optimize Content Based Image Retrieval system using Differential Evolution approach. In addition, it presents the concept of relevance feedback to reduce the semantic gap to improve the result. Besides this a complete background which is essential to understand and develop the concept of CBIR system is discussed.