At Model Health, we believe that cutting edge biomechanics has the power to transform physical health and boost human performance - but it needs to be accessible to every movement professional, not just research labs.
We spent years engineering the most portable lab-grade motion capture technology. We are now thriving to make it actionable to every practice and set up.
The app will walk you through three steps to pair the recording devices, calibrate them, and record activities .
Once the activity is recorded, the videos are uploaded to our cloud, where computer vision, AI, and biomechanics algorithms generate an accurate 3D biomechanical model.
The 3D model yields rich time-series data; including joint angles and angular velocities, global displacement, and center-of-mass trajectories.
We bridge the gap to real-world applications by extracting interpretable metrics and actionable insights specific to each movement from time-series data.
Drawing from scientific literature, we highlight key metrics that are linked to both performance optimization and injury risk -- empowering practitioners to make informed, evidence-based decisions and track patient or athlete progression.
OpenCap combines advances in computer vision, machine learning, and musculoskeletal simulation to make movement analysis widely available without specialized hardware, software, or expertise.
We validated OpenCap against laboratory-based measurements in a cohort of 10 individuals for a set of tasks including gait, squat, sit-to-stand, and drop vertical jump:
We also demonstrated OpenCap's usefulness for applications including screening for disease risk, evaluating intervention efficacy, and informing rehabilitation decisions.
We developed a more accurate and generalizable model, named marker enhancer, to predict the position of 43 anatomical markers from 20 keypoints identified from video. We trained this model on a large database of 1,433 hours of data from 1,176 subjects.
We showed the our model improves kinematic accuracy (4.1° error) compared to OpenCap's original model (5.3° error) on a benchmark dataset. We also showed that it better generalized to unseen, diverse movements (4.1° error) than OpenCap’s original model (40.4° error).
We developed a computationally efficient optimal control framework to predict human gaits based on optimization of a performance criterion without relying on experimental data.
The ability to predict the mechanics and energetics of a broad range of gaits with complex 3D musculoskeletal models allows testing novel hypotheses about gait control and hasten the development of optimal treatments for neuro-musculoskeletal disorders.
