combining generativediscriminative learning for automatic image annotation and retrieval结合generativediscriminative学习自动图像标注和检索.pdfVIP

International Journal of Intelligence Science, 2012, 2, 55-62 / 10.4236/ijis.2012.23008 Published Online July 2012 (http://www.SciRP.org/journal/ijis) Combining Generative/Discriminative Learning for Automatic Image Annotation and Retrieval 1 1 2 2 Zhixin Li , Zhenjun Tang , Weizhong Zhao , Zhiqing Li 1College of Computer Science and Information Technology, Guangxi Normal University, Guilin, China 2College of Information Engineering, Xiangtan University, Xiangtan, China Email: {lizx, zjtang}@, {zhaoweizhong, lizhiqingchina}@ Received April 24, 2012; revised May 30, 2012; accepted June 9, 2012 ABSTRACT 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 continu- ous probabilistic latent semantic analysis (PLSA) to model continuous quantity. Furthermore, we propose a hybrid framework which employs continuous PLSA to model visual features of images in generative learning stage and uses ensembles of classifier chains to classify the multi-label data in discriminative learning stage. Since the framework combines the advantages of generative and discriminative learning, it can predict semantic annotation precisely for un- seen images. Finally, we conduct a series of experiments on a standard Corel dataset. The experiment results show that our approach outperforms many state-of-the-art approaches. Keywords: Automatic Image Annotation; Continuous PLSA; Semantic Gap; Hybrid Approach; Image Retrieval 1. Introduction