Design of an experimental platform of gait analysis with ActiSense and StereoPi

Abstract Gait analysis is a systematic study of human movement. Combining wearable foot pressure sensors and machine learning (ML) solutions for high-fidelity body pose tracking from RGB video frames could reveal more insights into gait abnormalities. However, accurate detection heel strike (HS) toe-off (TO) events crucial to compute interpretable parameters. In this work, we present an experimental platform the timing using new sensor (ActiSense System, IEE S.A., Luxembourg), Google’s open-source ML solution MediaPipe Pose. For purpose, two StereoPi systems were built capture stereoscopic videos images in real time. Pose was applied synchronized cameras, algorithms (ALs) developed detect HS TO analysis. Preliminary results healthy subject walking on treadmill show mean relative deviation across all time spans less than 4% ActiSense device 16% AL2 (33% AL1) employing videos. Finally, work offers development sensor- video-based ALs automatically identify individuals those with disorders.