SkaFld.IgniteClinical / Sports Medicine

iMagine

Medical-grade 3D movement analysis for physical therapy — using standard cameras instead of a motion-capture rig.

The Problem

Physical therapy assessment relies on subjective visual observation — inconsistent between practitioners, slow, and impossible to track objectively over time.

What We Did

A computer-vision platform that captures real-time 3D movement from multiple cameras, runs AI pose detection, automatically scores functional movement patterns, and generates clinical reports with longitudinal progress tracking. We took a Python prototype to a full-stack clinical platform.

What We Built

Multi-Camera Capture

Synchronized real-time 3D capture from standard USB cameras.

AI Pose Detection

MediaPipe pose estimation with OpenCV stereo vision and 3D triangulation.

Automated Scoring

FMS protocols plus 24 exercise-specific assessments.

Clinical Reporting

Generated PDF reports with longitudinal progress tracking.

Results
52
API routes shipped
<100ms
Goal: real-time feedback latency
40%
Goal: reduction in assessment time
3D
Markerless Movement Analysis
Status

MVP complete (capture, live assessment, reporting). EMR integration in progress.

Stack & Approach
Next.js 15 · Three.jsFastAPIMediaPipe · OpenCVPostgreSQL · PrismaSocket.ioRailway