Pose estimation
Also known as: keypoint detection, skeleton tracking
A computer-vision technique that locates the body's joints in each frame of a video.
Why it matters
Pose estimation is the technology that makes phone-based gait analysis possible at all — understanding how it works helps you judge how much to trust the numbers it produces.
What it means
FormStride uses TensorFlow.js MoveNet to track 17 body keypoints (ankles, knees, hips, shoulders, elbows, wrists, ears, eyes, nose) per frame, then derives biomechanical metrics from their trajectories. Accuracy depends heavily on video quality: a clean side-on angle, steady frame rate, and good lighting all reduce noise in keypoint tracking and make the derived metrics more trustworthy.
Related reading
The complete guide to running form analysis