A team of researchers at the University of Notre Dame are working on a software that could identify criminals.

The technology is based on the thought that criminals are likely to visit their crime scene several times before or after a crime is committed. The U.S. military in the Middle East also indicated that IED bomb makers return to the locations of explosions to improve their designs.

Facial recognition that is not compared against an existing database can be used in a new technology to determine if certain people return to certain locations suspiciously often. It is up to law enforcement to determine what would be considered too often. Kevin Bowyer and Patrick Flynn of Notre Dame’s Computer Science and Engineering Department have been working on image-based biometrics since 2001. Together with Jeremiah Barr, a doctoral student in computer science and engineering, the team developed a “Questionable Observer Detector (QuOD)”, which records “face tracks” of individuals in videos and uses that data to identify people if they are appearing more than just one time.

Bowyer, Flynn and Barr admit that there are still “challenges” and there is work to do to overcome lighting issues as well as scenarios in which videos of faces are taken from different angles. However, they said they will be able to solve the problems and their software may provide law enforcement a new tool for identifying potential criminals.

SOURCE via University of Notre Dame