Reflections caught in the eyes of child sex abuse victims in photographs could help police to identify and catch the perpetrators, thanks to new research from two UK universities.
Psychologists at the University of York and the University of Glasgow were able to extract images of people standing behind the photographer from reflections in the eyes of subjects. They then discovered that they were identifiable by third parties, despite their low resolution and poor quality.
The iris typically occupies only 0.5 per cent of the area of a person’s face, meaning that the amount of reflected data that can be extracted to form an image is extremely small. Advances in camera technology have increased the number of pixels captured over the years but researchers found that the quality of images of reflected faces remains around 30,000 times lower than a face directly visible in the same photograph.
In the experiments, published in the journal PLOS One this month, people were photographed with a 39 megapixel camera and reflections in the eyes of the subject were cropped and magnified. The quality of camera used in the experiment is already available commercially; Nokia’s Lumia 1020 smartphone is already on the market with a more powerful41 megapixel sensor.
Previous research has shown that people can identify familiar faces from extremely low resolution images, down to just seven by ten pixels. Current technology means that it is possible to gather images of at least this quality. But the scientists were surprised to discover that even unfamiliar faces are distinguishable at this quality.
