Individual Image Retrieval Based on User Interest Model

There exists a semantic gap between low-level visual feature and high-level semantics feature, and the accuracy of image semantics annotation depends greatly on the description of low-level visual feature. Taking this into consideration, user interest model is proposed in this paper, syncretizing color feature and texture feature into eigenvector, labeling image semantics information by using user interest model. Experiments show that user interest model can be successfully used in image semantics annotation and individual image retrieval.