Samuel Yanqi Li ¹ and Tyler Huynh ², ¹USA, ²California State Polytechnic University, USA

The Lunge Flow app addresses the challenge of improving fencing techniques and preventing injuries by providing AI-driven, personalized feedback on user-uploaded videos [1]. Combining pose estimation, K-Means clustering, and ChatGPT, the app analyzes user movements and compares them to reference techniques. Experiments revealed an 80% accuracy rate for feedback matching professional evaluations and a high user satisfaction score of 4.6. The app’s key strengths are its accessibility, real-time feedback, and potential to enhance training outcomes. Future improvements include expanding technique coverage, refining visual design, and improving analysis for low-quality videos. Lunge Flow is a reliable, innovative tool for fencers and athletes aiming to perfect their craft.

Fencing, Pose Estimation, K-Means Clustering, Sports Training, Technique Improvement