Research and tutorials on human pose estimation.
Does single-leg jump strength predict soccer kick speed? In 6 amateur players, the dominant leg kicked harder (99 vs 85 km/h) — but the real surprise was that a stronger non-dominant support leg went with a faster dominant-leg kick.
Apple’s ARKit turns an iPhone into a markerless 3D motion tracker — but how accurate is it? I tested it across 10 exercises with 11 participants. Single joints at peak flexion were within ~3°, but errors near full extension and the floor reached up to 63°.
Can a smartphone app really measure how you move? Five motion-capture apps — ViMAS, KinesioCapture, Coach’s Eye, LGait, and SmartGait — have been validated against VICON and 3D systems. Here’s the accuracy each study actually found, app by app.
Can your phone measure how you move accurately enough to fix your form? Here’s why Apple’s ARKit — 3D motion capture from ordinary 2D video, auto-calibrated and real-time — is a promising answer, and how it compares with VICON, 2D video, and Kinect.
We benchmarked pose estimation accuracy across six models — MediaPipe, MeTRAbs, YOLO and MoveNet — against the VICON gold standard. Two clear tiers emerged; here’s which to use, and when.
Can markerless pose estimation replace marker-based motion capture in sports? Our validation study found a ~10-degree mean joint-angle difference — invisible to the eye, but too much for clinical precision. Here’s what that means.
A practical getting-started guide to two free markerless motion-capture tools: OpenPose (install the demo, fix the models, run it) and OpenCap (2 iOS cameras, calibration, and 3D capture).
Part 5 (finale): kinetic/dynamic simulation with OpenSim Moco and automatic differentiation – muscle activations and ground reaction forces from the reconstructed motion.
Part 4: driving an OpenSim model with the reconstructed motion – convert to .trc, LSTM marker augmentation (OpenCap), scale the model, and run inverse kinematics.
Part 3: triangulating 2D keypoints into 3D with the DLT, weighting by OpenPose confidence, person tracking, residual-analysis filtering, and reprojection-error optimization.