An Efficient Fuzzy Hashing Model for Image Retrieval

Image retrieval models based on visual perception have traditionally placed more emphasis on object shape than other image features. Indexing and retrieval techniques are two major concerns for efficient management of multimedia databases. In this paper, a new shape-based image retrieval model coupled with grey theory, a fuzzy hashing table and relevance feedback based on association rules is proposed to search for images with similar object shape. A GM(1,N) model is used for the shape-based image retrieval. Next, a fuzzy hashing scheme is employed to solve the problem of hash misses. Furthermore, association rules mined from users' retrieval history are used to reveal users' image searching behavior. The proposed strategy is applied to the FISH dataset to demonstrate its effectiveness