Abstract
This paper uses content-based image retrieval technology and presents a reasoning-based image semantic indexing and retrieval approach, which could bridge the gap between complex structure of shapes and the extracted low-level features of images. It starts with image features extraction as other content-based image retrieval approaches do. Then applies structured knowledge representation methods to index, classify and recognize objects automatically according to the structured and semantic attributes of them. Syntactic and semantic descriptions of this indexing and retrieval model and some reasoning services involved in the issue are also analyzed and discussed. Finally, a simple semantic indexing and retrieval model framework of this method has been presented.