Robert Collins

Dr. Collins joined the faculty of CSE as an associate professor in spring 2005. Prior to joining CSE, he was an associate research professor at the Robotics Institute of Carnegie Mellon University. He received his Ph.D. in computer science in 1993 from the University of Massachusetts in Amherst, MA, for work on scene reconstruction using stochastic projective geometry.

Dr. Collins is co-director of the Laboratory for Perception, Action, and Cognition (LPAC) in the CSE department. His research area is computer vision, with an emphasis on video scene understanding, automated surveillance, human activity modeling, and real-time tracking. From 1992 to 1996, he was technical director of the DARPA RADIUS project at UMass, which developed vision algorithms for recovering 3D site models from multiple, oblique aerial views. From 1996 to 1999, Dr. Collins was technical director of the DARPA Video Surveillance and Monitoring (VSAM) project at CMU, which demonstrated a real-time, automated multi-camera video surveillance system for monitoring the activities of people and vehicles in a complex scene. Dr. Collins developed video visualization tools and rapid camera calibration routines for the 30-camera EyeVision system, demonstrated during the live broadcast of Superbowl XXXV in January 2001. From 2002 to 2004, he was co-PI of the DARPA HumanID project, which explored new algorithms for biometric identification based on gait, non-frontal facial views, and facial expression. His current research is on real-time, appearance-based object tracking, with a special emphasis on tracking moving objects from moving camera platforms. This work has been used to perform surveillance from unmanned air vehicles (DARPA VIVID program, Co-PI) and to navigate an unmanned water vehicle in river and harbor environments (DARPA Mars2020 program, PI). Dr. Collins is PI of a current NSF grant on persistent tracking of objects in video and a Co-PI of a multi-disciplinary NSF grant for applying computer vision technology to analyze the social behavior of individuals and groups in crowds.  He is an associate  editor for the International Journal of Computer Vision.

Research Interests:

Computer Vision, with current emphasis on Video Scene Understanding, Automated Surveillance, Human Activity Modeling, and Real-time Tracking

Selected Publications:

1. Park, M., K. Brocklehurst, R. Collins, Y. Liu.  2009.  Deformed Lattice Detection in Real-World Images.  To appear in IEEE Transactions on Pattern Analysis and Machine Intelligence, Special Issue on Probabilistic Graphical Models.
   
2. Ge, W., R. Collins.  June 2009.  Marked Point Processes for Crowd Counting.  To appear in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2009).  8 pages.  Miami, FL.
3. Yin, Z., R. Collins.  June 2009.  Shape Constrained Figure-Ground Segmentation and Tracking.   To appear in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2009).  8 pages.  Miami, FL.
4. Collins, R. T., Y. Liu, M. Leordeanu. October 2005. On-Line Selection of Discriminative Tracking Features. IEEE Transaction on Pattern Analysis and Machine Intelligence 27(10):1631-1643.
   
5. Collins, R. T., A. Lipton, H. Fujiyoshi, T.Kanade. October 2001. Algorithms for Cooperative Multi-Sensor Surveillance. Proceedings of the IEEE 89(10):1456-1477.