Web image co-clustering based on tag and image content fusion

In Web 2.0 applications, users always label digital images using textual descriptions, which are also called tags. As a result, a web image usually carries both tag and visual content information. In order to improve the retrieval performance of web images, in this paper, we propose an error-driven fusion co-clustering algorithm, which combines images´ tags, visual contents together for analysis. Experimental results demonstrate that our algorithm outperforms other simple clustering methods.