Core clinical engineering and protocols for Video Gait Analysis.
Traditional biomechanical gait analysis requires dedicated multi-camera optical labs with reflective skin markers, costing tens of thousands of dollars and confining evaluations to artificial laboratory environments.
HealthRecognition utilizes markerless pose estimation networks running on everyday smartphone video. Patients simply walk past the phone camera, and the algorithm extracts hip, knee, and ankle joint angles, calculating stride length, velocity, and arm swing asymmetries to objectively evaluate Parkinson's disease progression.