Image Annotation and Retrieval Based on Bayesian Theory

Automatic semantic annotation of image is a crucial and highly challenging work in content-based image retrieval.A method of constrained clustering is proposed to cluster the segmented region,during the annotation stage,a greedy selection and joining method is used to find out the independent region cluster subsets.Then,a method of Bayesian is used to image annotation.After the image annotation,we can use the annotation Keywords to image retrieval;the system has been implemented and tested on an image database of about 500 images.The experiment results show the effectiveness of the proposed approach.