|
|
Jamie Sherrah, Shaogang Gong, A.Jonathan Howell, Hilary Buxton
While full computer understanding of dynamic visual scenes containing several people may be currently unattainable, we propose a computationally efficient approach to determine areas of interest in such scenes. To this end, we present methods for modelling and interpretation of single- and multi-person human behaviour in real time to control video cameras for visually mediated interaction. We demonstrate that while environments containing a single person are relatively simple, interpretation of multi-person scenarios is much more difficult.