Researchers develop a new technique to identify faces from reflections

Researchers conducted an experiment where image of bystanders were extracted from the corneal reflections of the photo images of an individual.

Published Date
30 - Dec - 2013
| Last Updated
30 - Dec - 2013
Researchers develop a new technique to identify faces from reflec...

Scientists have developed a new forensics technique in which they were able to extract images from the corneal reflections of an individual of people standing nearby. The technique can be useful while conducting criminal investigations.

Researchers Rob Jenkins from the University of York in Britain, and collaborator Christie Kerr, from the University of Glasgow, conducted the experiment in which they found that faces can be reconstructed from images taken of an individual with commercial digital cameras and enhanced with off-the-shelf image-processing software.

For the experiment, the scientists used a high-end digital camera and sat five volunteers for a passport-photo-like shot, using studio lighting. All the five people where in the room even when not being clicked and stood close to the photographer to be included in the reflection off the subject's corneas.

Then they asked the subjects to identify images from the reflections as well as the original digital images. Nine out of 10 volunteers were able to correctly identify from the corneal images. The group which was familiar with the five subjects were able to identify the persons with 84 percent success rate and those unfamiliar with the subjects in the photographs were able to identify with 71 percent success rate.

The researchers added that this forensics technology can be used in crimes where the victims are photographed, like abductions or child abuse. The reflections in the eyes of the photographic subject could help in identifying perpetrators. They added that the images of people retrieved from cameras seized as evidence during criminal investigations may be used to piece together networks of associates or to link individuals to particular locations.

"The pupil of the eye is like a black mirror. To enhance the image, you have to zoom in and adjust the contrast. A face image that is recovered from a reflection in the subject's eye is about 30,000 times smaller than the subject's face. Our findings thus highlight the remarkable robustness of human face recognition, as well as the untapped potential of high-resolution photography," Jenkins said.

"You could think of it as a foray into extreme facial recognition. Yes, the camera can resolve the face, and yes, the brain can identify it," he writes, "but both systems are pushed to their limits, and neither could perform the feat alone."

Source: Plos One