An image retrieval method of getting the high-level semantics from the low-level features

This paper presents results of a project that seeks to transform the low level features to high level of meaning. It extracts the low level features called as the representative colors from the images, and presents a new approach called word net to establish the link from the low-level feature vectors to the semantics. In order to improve the retrieval efficiency, the relevance feedback is also applied into the system. Experiments show the method is promising.