Great deal images are available by the rapidly developing of computer, multimedia, and Internet techniques. It is an important and challenging research topic that rapid and effective searching for desired images from large-scale image databases. Content-based image retrieval (CBIR) is the set of techniques to address the problem of retrieving relevant images from an image database based on automatically derived image features. In recent years, CBIR is a very active research direction and has been applied to many fields.The main researching work of this dissertation has been done around the low-level feature extractions. Firstly, this paper present a novel color image retrieval algorithm based on the theory of bit-plane, which is using features of the color and the spatial distribution of color. Secondly, three new concepts are defined, which are bit-plane mean, bit-plane flatness and bit-plane roughness, in order to describe the retrieval features perfectly and to improve the retrieval efficiently and according to these three concepts, the features of color and the spatial distribution of color are extracted. Finally, a reasonable measure method of similarity is designed with the characteristics of these retrieval features. Experiments show that this novel color image retrieval algorithm is effectively and accurately in retrieving the similar images. |