Barbell Trajectory Tracking for Performance Analysis During Snatch Movement in Weightlifting

Dhairya Shah, Christopher Taber, Tolga Kaya, Eva Maddox, Mehul S. Raval

Abstract: Olympic-style weightlifting involves complex and technical movements where accurate tracking of barbell motion is crucial for performance analysis. In this paper, we present a computer vision based framework that first corrects for perspective distortion caused by varying camera height and distance, then employs a rule-based algorithm to classify snatch trajectories into four distinct types. Preliminary investigation on 6000 frames suggests 70% classification accuracy. Building on these labels, eight key barbell kinematic variables were calculated and utilized three—vertical peak height (\(Y_{\text {max}}\)), initial horizontal setup (\(X_{\text {1}}\)), and bar drop efficiency (\(Y_{\text {catch}}\)) to generate a consolidated 0–4 performance score, mapped to five qualitative categories from “Very Bad” to “Excellent”. This two fold approach, comprising trajectory classification and score calculation, was validated by a sports scientist, ensuring its reliability in helping athletes optimize lifting techniques by providing insights into barbell trajectory patterns.

External IDs:dblp:conf/isace/ShahTKMR25