Show HN: I made a computer vision tool for evaluating deadlift form

Category: ai-ml

Tags: computer-vision, pose-estimation, fitness, video-analysis, pytorch

Score: 6.0/10 (Innovation: 6, Technical: 7, Documentation: 5, Utility: 6)

A collection of computer vision demos that use pose estimation and segmentation models to analyze athletic movements like deadlifts, pull-ups, and rock climbing. The deadlift demo counts reps and classifies back posture as straight or rounded, combining SAM3.1, VitPose+, and Gemma for a multi-model pipeline. It's interesting as a practical application of vision models to fitness form feedback.

Target audience: fitness enthusiasts, computer vision developers, sports scientists

Repository: https://github.com/jeremyipark/vision-demos/tree/main/deadlift · Python · Apache-2.0 · 857 stars

View on Hacker News