Face Recognition May be Better with a Body: Computer Identity Programs Could Benefit

Programming computers to recognize people usually means focusing exclusively on faces. Now, though, researchers have found that it may be beneficial to also focus on a person's body. It turns out that there is potentially more valuable information for biometrics-based identity recognition in images of people than the face alone.

"For twenty years, the assumption in the automatic face recognition community has been that all important identity information is in the fact," said Jonathon Phillips, one of the researchers, in a news release. "These results should point us toward exploring new ways to improve automatic recognition systems by incorporating information about the body beyond the face."

This particular study involved a series of experiments. These experiments showed study participants pairs of images of either the same or different people and asked them to determine if the photos matched or not. The images were chosen from a database used in the Face Recognition Vendor Test 2006, an international competition of face recognition systems conducted by NIST.

The images themselves included a subject's face and their upper body. Surprisingly, the researchers found that biometric identification accuracy by humans was essentially random when viewing only the face. When showed the upper body, though, participants were far more accurate. This seemed to show that the volunteers primarily made decisions based on the upper body.

The researchers also tracked the eye movements of the participants. Although participants presented with limited facial information reported using facial cues to make identifications, their eye movements told a different story.

"Eye movements revealed a highly efficient and adaptive strategy for finding the most useful identify information in any given image of a person," said Alice O'Toole, one of the researchers, in a news release.

The findings reveal that taking the body and not just the face into account in automatic recognition systems may be important. More specifically, it has the potential to improve accuracy by incorporating the body into the decision process.

The findings are published in the journal Psychological Science.