Face Averaging Technique

Face Averaging Technique

CCTV (Closed-Circuit Television) records are critical in forensic examinations. It can easily provide evidence to law enforcers as to what really transpired in a crime scene. However, not all CCTVs are equipped with state-of-the-art sensors and lenses. As a result, suspects caught on tape are not recognizable, even after implementing the different techniques that are currently available.

This is the problem that psychologists from the universities of Lincoln, New York, and New South Wales collaborated on. The researchers tested combining a series of pictures using a ‘face averaging’ technique — a method which digitally combines multiple images into a single enhanced image, removing head angles or lighting, so that only the recognizable traits to indicate the identity of a person remains. Their theory was that it would improve the overall quality of the image that would be captured from different low-quality CCTV footages.

The researchers tested their study by having humans and a computer facial recognition software identify people from quality images, pixelated images, and ‘face averages’. The test results showed that humans and the software were better at identifying someone when viewing the average image that is combined from multiple pixelated images than a single quality image. In some cases, the computer recognition software reached a 100 percent accuracy when identifying someone using the ‘face averaging’ technique.

This innovation has huge impacts on law enforcement agencies. The team leader of the research, Dr. Kay Ritchie, mentioned that they have shown a “relatively quick and easy way to improve pixelated images of someone’s face”. Their algorithm could greatly help in identifying a suspect from low-quality CCTV footages in which there are different cameras available. The researchers even tested their method by running it through different computer facial recognition software to prove that their method works as expected — one is by using a smartphone app, and the other was the one widely used by forensic scientists. In both cases, by using the aforementioned method, the chances of identifying a person was higher as compared to using a single image, thus, proving their theory.

Want to know more about this innovation? Just follow this link!

Like Us on Facebook!

- Fun Facts and Trivia brought to you by the Research and Development Committee -

Thank you and God Bless!