Video Surveillance and Monitoring is very active area of research in Computer Vision. However, most of the current approaches assume that the observed scene is not crowded, and that reliable tracks of objects are available over longer durations. Therefore, these approaches are not extendable to more challenging surveillance videos of crowded environments like markets, subways, religious festivals, parades, concerts, football matches etc, where tracking of individual objects is very hard, if not impossible.

In this talk, first I will present an approach for tracking people in crowded scenes using multiple cameras. Our approach uses a homographic occupancy constraint (HOC), which states that if a foreground pixel has a piercing point that is occupied by foreground object, then the pixel warps to foreground regions in every view under homographies induced by the reference plane, in effect using cameras as occupancy detectors. Using HOC we are able to resolve occlusions and robustly determine locations on the ground plane corresponding to the feet of the people, and track them in subsequent frames.

Next, I will present a framework for modeling scenes involving high density crowds in which Lagrangian particle dynamics are used to segment crowd flows and detect any flow instability. For this purpose, flow fields generated by moving crowds are treated as an aperiodic dynamical system which is manifested in terms of time dependent optical flow. A grid of particles is overlaid on the flow field, and particles are advected using a numerical integration scheme. This is followed by the quantification of the amount by which the neighboring particles have diverged using a Cauchy-Green deformation tensor.

Finally, I will discuss an algorithm that tracks an individual within the crowd. The approach is based on the observation that a pedestrian behavior in crowds results from the collective behavioral patterns evolving from the space time interaction of large number of individuals among themselves and with the geometry of the scene. Therefore, we incorporate the influences generated by other individuals of the crowd and scene geometry into the tracking algorithm itself.