Researchers have developed a deep learning-based approach to estimate the severity of Parkinson's Disease by analyzing how patients turn around in their homes. The method uses videos taken from a single camera and measures the rotation of hip and knee joints. It was tested on two datasets, one collected from people with PD and another from healthy volunteers, achieving an accuracy of 41.6% and a mean absolute error of 34.7 degrees. This work focuses on free-living home settings, where complexities like baggy clothing and poor lighting can affect measurements. --- Why it matters: This research matters to engineers because it aims to develop a passive and continuous method for assessing PD severity in real-world settings, which could improve diagnosis and treatment outcomes. Source: https://arxiv.org/abs/2408.08182