Kingston University researchers are developing techniques to help security personnel and the police to quickly analyze CCTV footage of criminal incidents.
Dr James Orwell, Kingston University London
Visual surveillance experts belonging to the Digital Imaging Research Centre at the university are working on a novel project that could aid in catching looters and rioters and bring them to justice. The three-year project commenced from February 1, 2011.
Dr James Orwell at the Kingston University said that the project aimed to develop components for automatically analyzing CCTV footage of trigger incidents from multi-camera networks and then reproduce parts of the video footage pertaining to the investigation. He explained that in an incident where a looter breaks into a shop, the technology would enable tracing of the activities of the looter from when the person entered the specific area and also track the activities of the person after the incident. The looter may lower his or her guard and remove the face mask and could be identified.
The research aims to tackle issues related to video surveillance, which data is to be stored and which personnel should be granted access to the data. An intruder-detection system could help identify other video data that contain footage of the looter’s activities. New technology for identifying ‘triggers incidents’, such as suspicious behavior, fights, or riots are to be developed. It would also address human rights issues relating to privacy, freedom and equality.
Current techniques that use face-recognition technologies are not suitable for public video surveillance. Their image quality is inadequate for identifying faces. The project aims to develop alternative techniques to address this problem.
Kingston researchers aim to develop new techniques for data indexing, storage and management. It will allow personnel to view only relevant data and also help reduce infringement of privacy. The project would also develop techniques for securely erasing impertinent data.