An RGB-D sensor-based instrument for sitting balance assessment

Sitting balance is an important aspect of overall motor control, particularly for individuals who are not able to stand. Typical clinical assessment methods for sitting balance rely on human observation, making them subjective, imprecise, and sometimes time-consuming. The primary objective of this study is to develop an improved system for assessing sitting balance in clinical settings. We designed a software tool that takes input from an RGB-D camera system to track human movement during sitting balance assessment. We validated the system by tracking subject's movements during two seated balance exercises. To assess the accuracy of our system's measurements, we compared them with measurements taken using a ruler and measurements captured from still images from a video recording. The agreement of body angle measurement was an average of 2.19 +/- 2.29 degrees, and agreement of forward reach distance was an average of 0.1 +/- 0.25 in.. The results show that our approach can track a person's body movements with clinically relevant accuracy, suggesting that this RGB-D camera-based system could offer advantages over existing observational methods of sitting balance assessment.