Markerless Behavior Monitoring System for Diagnosis Support of Developmental Disorder Symptoms in Children

This study presents a markerless behavior evaluation system employing multiple RGB cameras and Kinect V2 sensors to assists clinicians in identifying disorder symptoms in children. The system utilizes OpenPTrack with Kinect sensors to track children’s and toys’ positions and records their activity using RGB cameras. Children’s activity was estimated by computing the distance between them and the toys. Children’s behavior was modeled with a Petri net, and four features were extracted from the model. We conducted preliminary experiments with four typical and three ASD disorder children. The experimental results demonstrated that the frequency of changing activity and playing alone was more informative than the others to distinguish ASD children from the typical ones.

Recommended citation: Putra, P.U., Shima, K., Hotchi, S., & Shimatani, K. (2021). Markerless behavior monitoring system. *ICCAS 2021*.