Novel Technology Detects Lies Through Movements Of Facial Muscles

Dec, 2021 - By SMI

Scientists developed a sensing technology over 15 years that exposes liars through contorting muscles linked to deception, and may lead to software and cameras that detect these facial movements by themselves.

Deception could be observed everywhere in life, from social interactions to affairs of homeland security. However, reliable indicators for such deceptive behaviors are still indefinable. Now scientists form Tel Aviv University developed a lie-detecting system which according to the scientists outperforms any other method. Over the 15 years, scientists developed this system that uses electrodes attached to the face in order to expose contortion of the muscles that are linked to deception.

In this study, the scientists studied how electrodes commercialized by An Israeli spin-off venture known as X-trodes, could be useful for revealing deception in people. The team used these electrodes to measure contortions of the facial muscles related to deception. The scientist paired the subjects and suggested them to sit opposite to each other. One of the subjects from each pair was given headphones that played word either “tree” or “line” and were told to convey the word to the partner deceptively or truthfully. As the partners couldn’t detect the lies with any reliability, the new sensing system successfully detected the lies.  The system included electrode stickers that were placed on two facial muscles groups, the muscles above the eyebrows and the cheek muscles near the cheeks. The systems passed electrical signals to a machine learning algorithm that was trained to differentiate truths from lies.

Furthermore, the algorithm detected lies of subjects with significant rate of 73% success. According to the scientists, the study shows that wearable electrodes can be utilized in detection of lies, moreover, it envisions with proper machine learning algorithms and video cameras, a system could be built that detects deception simply by analyzing high-res imagery.