Researchers Create Software That Could Identify Fake Facial Expressions

Fake and real smiles could be challenging to differentiate. However, researchers have come up with new computer software which can identify false facial expressions. According to a published study, this software will record the smile movement on the face of a person. Then it will analyze the data to determine whether that expression is true or not.

Moreover, the software also detects the most significant muscle movements in the area around the person's eyes. That supports the theory that you can see a genuine and spontaneous smile in one's eyes.

Why do we need it?

Hassan Ugail, the lead author of the study (Bradford University's Professor) said that in the last couple of years, techniques to analyze people facial expressions have progressed dramatically. However, differentiating between false and genuine facial expressions remains difficult since we are not good at picking up relevant cues.

This software functions by using video recording to map and identify the cheeks, eyes, and mouth on the subject's face. Then, it will measure how these features will change with a smile. After that, it calculates the movement differences between clips of fake and real smiles.

How did they test this software?

Researchers have used two datasets to test this program. One set has images of people with genuine smiles, while the other one has images of posed smiles. Comparing between the fake and real expressions, they found significant differences in the movement of cheeks and mouths of the subjects. However, the most relevant variation is the eyes area where genuine smile generates 10% more of muscle movements.

When we fake a smile, often only the muscles in the mouth area move, however, we usually couldn't spot the difference in the eyes area. This new computer software can help us do that.

Hassan Ugail said that this new technology could be used to improve the interactions between humans and computers - like with biometric identification.