New facial, gait recognition software to be integrated in CCTVs

Published 25 February 2009

BAE Systems and OmniPerception work on developing a gait and facial behavior recognition to be integrated into street corner CCTVs; it will make identifying known criminals easier

Nowhere to hide — almost. Identifying criminal suspects on CCTV could soon become easier and more accurate following a £1.5 million collaboration between BAE Systems and Guildford, Surrey, United Kingdom-based facial recognition expert, OmniPerception. The joint project aims to improve the use of surveillance systems by enhancing existing biometric software with gait recognition and facial behavior analysis. The research will also explore advanced automatic tracking technology that could be used during large-scale public events.

David McIntosh, chief executive of Surrey-based OmniPerception, said: “The nation has spent an awful lot of money on cameras. We have more CCTV cameras per head than any other country in the world, with at least half a million of them in London alone. Although these cameras are taking images all the time, the footage is often unusable. What we’re trying to do with BAE is make CCTV and other imaging systems a lot more intelligent. We not only want to monitor facial features, but also automatically track distinctive characteristics such as the way someone walks or smiles. Raw images are some of the most difficult things for computers to identify, and this is something that we hope to improve.”

The project is one of the first to receive funding as part of the BAE Systems Investment in Innovation program. In addition to assistance from BAE, the research will gain support from Surrey University’s Center for Vision Speech and Signal Processing (CVSSP), which will receive around £250,000 of the total funding. McIntosh expects to roll out the system to police forces across Britain within the next eighteen months.