Academic
Publications
Stream-Based Active Unusual Event Detection

Stream-Based Active Unusual Event Detection,10.1007/978-3-642-19315-6_13,Chen Change Loy,Tao Xiang,Shaogang Gong

Stream-Based Active Unusual Event Detection   (Citations: 1)
BibTex | RIS | RefWorks Download
We present a new active learning approach to incorporate human feedback for on-line unusual event detection. In contrast to most existing unsupervised methods that perform passive mining for unusual events, our approach automatically requests supervision for critical points to resolve ambiguities of interest, leading to more robust and accurate detection on subtle unusual events. The active learning strategy is formulated as a stream-based solution, i.e. it makes decision on-the-fly on whether to query for labels. It adaptively combines multiple active learning criteria to achieve (i) quick discovery of unknown event classes and (ii) refinement of classification boundary. Experimental results on busy public space videos show that with minimal human supervision, our approach outperforms existing supervised and unsupervised learning strategies in identifying unusual events. In addition, better performance is achieved by using adaptive multi-criteria approach compared to existing single criterion and multi-criteria active learning strategies.
Conference: Asian Conference on Computer Vision - ACCV , pp. 161-175, 2010
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
Sort by: