Big Brother Biometrics

“Big Brother is watching.” It’s a phrase that will make even the most innocent person scan the surroundings for a surveillance camera. The connotation is always negative. But the truth is that video is a crucial tool for the good guys — those who serve and protect — whether the footage is used to identify potential terrorists, capture thieves moving between buildings, or detect troublemakers on a university campus. Unfortunately, the quality of the video, the lighting conditions, and the distance a subject is from a camera affect how well the footage can help law enforcement identify subjects and track their movements.

Kevin W. Bowyer, the Schubmehl-Prein Chair of the Department of Computer Science and Engineering, along with Professor Patrick J. Flynn and Research Assistant Professor Nitesh V. Chawla, are co-investigators on a project dealing with videographic biometric technologies, which is being funded by the Department of Justice. The goal of their project, titled “Face Recognition from Video,” is to study the effects of using a variety of video sources (which feature varying levels of quality) and to develop algorithms which will select and compare multiple frames from several video streams — the type acquired from surveillance applications — in order to maximize recognition levels.