Transcript link2000
DTI Management of Information LINK Project: ICONS Incident reCOgnitioN for surveillance and Security funded by DTI, EPSRC, Home Office (March 2000 - March 2003) Queen Mary Westfield College, University of London Safehouse Technologies Heritage Protection British Telecom British Airport Authority The Central Research Issue (Pre-competitive): Recognition, Searching and Categorising Spatio-Temporal Events of “Significance” (Smart Content) in Video Data (on- or off-line) • “Zero Motion” Detection – Detecting and searching pre-defined significant events (spatial and temporal) in a known scene • “Abnormal Behaviour” Recognition – Recognition and indexing human activities and temporal signatures in space and over time Robust Detection and Tracking of Multiple Subjects and their Faces in Real-Time • In Real-Time with limited resources (on PC) • From all different views • Varying lighting, distance and racial colours • Under partial and total occlusion • Multiple subjects of interest • Fault tolerant to missing frames Behaviour Profiling and Categorisation Tracking human presence Modelling patterns of activities Understanding scene layout Learning behaviour patterns Behaviour Recognition and Search Automatic Recognition of Human Gestures for Visually Mediated Interaction and Intelligent Human-Machine Interface Other Currently Funded Research Projects • Gesture Recognition in Visually Mediated Interaction (EPSRC) – Gong, Sherrah • Advanced Incident Monitoring System (Industry) – Gong, Parkinson, Verity, Ong • Learning Human Models from Live Video (British Council) – Ong, Gong • Motion-based Recognition of Face and Behaviour (HEFCE) – Psarrou, Gong, Romdhani, Walter • Tracking and Segmentation in Virtual Studios (BBC & EPSRC) – Raja, Gong, Paker • Multi-Camera Data Fusion (Overseas) – Chang, Gong • Visual Learning in Dynamic Scenes (University of London) – Ng, Lee, Gong Key References • • • • • • • S. Gong, S. McKenna, A. Psarrou. Dynamic Vision: From Images to Face Recognition. World Scientific Publishing & Imperial College Press, March 2000. S. Gong, A. Psarrou, S. Romdhani. Multi-view Face Appearance Models. Invited article (1999 British Machine Vision Conference Best Scientific Paper Award), Image and Vision Computing, July 2000. S. Gong, A. Psarrou and M. Walter. Learning Stochastic Models of Human Behaviour. Invited article, Image and Vision Computing, July 2000 (also in ICCV’99, September 1999, Corfu, Greece). E. Ong, S. Gong. Modelling Nonlinear Dynamics of View Transformation: An Application to Tracking 3D Human Skeleton Models in Multiple Views. Invited article, Computer Vision and Image Understanding, October 2000 (also in IEEE Int. Workshop on Modelling People, 1999). S. Gong. Understanding head pose and body gestures for visually mediated interaction. Invited talk in IEE Colloquium on Computer Vision for Virtual Human Modelling. 1998, London (also in ECCV’98, Freiburg, Germany). S. McKenna, S. Gong. Recognising moving faces. In Wechsler, Philips and Bruce (Eds.) Face Recognition: From Theory to Applications, NATO ASI Series F, Springer-Verlag, July 1998. H. Buxton, S. Gong. Visual surveillance in a dynamic and uncertain world. Artificial Intelligence, Invited special volume on computer vision, Vol. 78, October 1995 (also in ICCV’95, June 1995, MIT, USA).