Researchers at the University of North Carolina at Chapel Hill and the University of Maryland at College Park recently developed a deep learning software that can determine individuals moods based on their gait. This is done by analyzing a video of someone walking, converting it to a 3D model, and extracting their gait.
Once the dominant motions are determined, the software then cross-references its database to determine how the person feels. The deep learning software is able to narrow down one's emotion to four moods: happy, sad, angry, and neutral, and the AI technology boasts over 80% accuracy.
The research team hopes to be able to use deep learning to detect the public's emotion from runs, gestures, and to narrow detection into subtle emotions.
Image Credit: Shutterstock