IEEE Project Abstract

Accurately clustering Internet-scale Internet users into multiple communities according to their aesthetic styles is a useful technique in image modeling and data mining. In this work, we present a novel partially-supervised model which seeks a sparse representation to capture photo aesthetics 1. It optimally fuzes multi-channel features, i.e., human gaze behavior,quality scores, and semantic tags, each of which could be absent.Afterward, by leveraging the KL-divergence to distinguish the aesthetic distributions between photo sets, a large-scale graph is constructed to describe the aesthetic correlations between users.Finally, a dense sub graph mining algorithm which intrinsically supports outliers (i.e., unique users not belong to any community)is adopted to detect aesthetic communities. Comprehensive experimental results on a million-scale image set crawled from Flickr have demonstrated the superiority of our method. As aby product, the discovered aesthetic communities can enhance photo re targeting and video summarization substantially.