Researchers led by Penn State have developed a human movement dataset that aims to enhance artificial intelligence’s capability to identify emotions through body language. James Wang, a distinguished professor at the College of Information Systems and Technology, highlighted the importance of body movements in conveying emotions. The research focused on identifying specific motor patterns, or “motor elements,” associated with expressed emotions. The team used 1,600 video clips annotated with Laban Movement Analysis (LMA) to create the dataset. Chenyan Wu, a doctoral student, then designed a movement analysis network that utilizes this dataset to predict both emotion and LMA labels for new visual content.
https://techxplore.com/news/2023-10-human-body-movements-enable-automated.html