This report summarizes the accomplishments and research design methods of a project funded by the National Institute of Justice to develop an automated facial recognition system for use by law enforcement in existing closed-circuit television (CCTV).
The project developed a surveillance system that uses real-time face recognition technology to increase the usefulness of currently existing CCTV-compatible surveillance software. Two approaches significantly improved the performance of the best existing surveillance system. These approaches involved the development of techniques for (1) the dynamic adjustment of video parameters in the region of the image containing a face, and (2) tracking a face to acquire multiple images of it across video frames. The dynamic adjustment of video parameters involved automatic evaluation of image quality, compensation for image characteristics that were suboptimal in the original image but that could be improved using known image processing algorithms, and real-time adjustment of imaging parameters via a feedback mechanism to the camera. Tracking a face across multiple video frames allowed for better performance in terms of recognition, based on statistical information gleaned from multiple matching operations. The project outcome is a state-of-the-art, automated facial recognition surveillance system capable of being extremely useful to law enforcement, intelligence personnel, and CCTV control room officers. The automation of surveillance lowers overhead and frees resources for other tasks. Figures, footnotes, and illustration
Downloads
Similar Publications
- Research in Brief: Three-City Evaluation of Gunshot Detection Technology
- Space-Time Association between Gunshot Detection Alerts, Calls for Service, and Police Enforcement in Chicago: Differences Across Citizen Race and Incident Type
- Characteristics and Dynamics of Cyberstalking Victimization Among Juveniles and Young Adults