Feature-based tracking of multiple people for intelligent video surveillance.
Date of Award
CC BY-NC-ND 4.0
Intelligent video surveillance is the process of performing surveillance task automatically by a computer vision system. It involves detecting and tracking people in the video sequence and understanding their behavior. This thesis addresses the problem of detecting and tracking multiple moving people with unknown background. We have proposed a feature-based framework for tracking, which requires feature extraction and feature matching. We have considered color, size, blob bounding box and motion information as features of people. In our feature-based tracking system, we have proposed to use Pearson correlation coefficient for matching feature-vector with temporal templates. The occlusion problem has been solved by histogram backprojection. Our tracking system is fast and free from assumptions about human structure. We have implemented our tracking system using Visual C++ and OpenCV and tested on real-world images and videos. Experimental results suggest that our tracking system achieved good accuracy and can process videos in 10-15 fps.Dept. of Computer Science. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis2006 .A42. Source: Masters Abstracts International, Volume: 45-01, page: 0347. Thesis (M.Sc.)--University of Windsor (Canada), 2006.
Ali, Mohammad Ahsan., "Feature-based tracking of multiple people for intelligent video surveillance." (2006). Electronic Theses and Dissertations. 2089.