论文部分内容阅读
In order to bridge the semantic gap exists in image retrieval, this paper propose an approach combining generative and discriminative learning to accomplish the task of automatic image annotation and retrieval. We firstly present continuous probabilistic