A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science, School of Engineering and Technology

Thesis (Ph.D.) - Asian Institute of Technology, 2013

As public concern about crime and terrorist activity increases, the importance of public security is growing, and video surveillance systems are increasingly widespread tools for monitoring, management, and law enforcement in public areas. At the same time, video surveillance systems have become a popular research area in computer vision. Many algorithms exist to detect and track people in video streams. However, human detection and tracking in high density crowds, where object occlusion is very high, is still an unsolved problem. Pre-processing techniques such as background subtraction fail in such situations. The emphasis in this work is on the development of a fully automatic algorithm to detect and track multiple humans in high-density crowds in the presence of extreme occlusion. Typical approaches such as background modeling and body part-based pedestrian detection fail when most of the scene is in motion and most body parts of most of the pedestrians are occluded. Under these conditions, we believe that the head is the only body part that can be robustly detected and tracked. In this dissertation, we therefore present a method for tracking pedestrians that detects and tracks heads rather than full bodies. First, we use a Viola and Jones AdaBoost detection cascade to detect human heads in images. However, we do not rely particularly on the Viola and Jones method. Our method will improve upon the raw results of any head detector using a blend of detection, appearance-based tracking, and incremental head plane estimation. Any effective head detector could be plugged into our algorithm if it is capable of detecting heads in crowds.