International Journal of Advanced Research in Computer Science and Software Engineering (IJARCSSE)
The bottleneck of current Content Based Image Retrieval (CBIR) systems is the semantic gap between low level image features and high level concepts. In order to overcome this bottleneck, the most of the recent research work in CBIR is focused on reduction of semantic gap between user and system. This paper presents a comprehensive survey on various methods proposed in literature for the Semantic CBIR. The state-of-the-art techniques available in the literature are divided into three categories: relevance feedback techniques to integrate user's perception, machine learning techniques to associate low level features with high level concepts and machine learning using neural network.