Vivienne Zhai¹ and Andrew Park², ¹USA, ²California State Polytechnic University, USA

Golf instruction remains economically inaccessible for most recreational players, with professional coaching costing $50-$200 per hour. This research presents Twelfth Tee, an intelligent mobile application leveraging computer vision and artificial intelligence to democratize access to professional-grade golf swing analysis. The system employs a three-tier architecture combining Flutter-based mobile video capture, Python backend pose estimation processing, and OpenAI GPT-4 integration for personalized feedback generation [1]. The methodology utilizes deep learning-based human pose estimation to extract 33 body landmarks from smartphone-recorded golf swing videos, calculating biomechanical metrics including joint angles, swing phases, and movement patterns. These quantitative measurements inform GPT-4 prompts that generate contextual coaching advice tailored to specific techniques and detected issues. Experimental validation demonstrated pose estimation accuracy within 4.2-6.8° mean absolute error across camera perspectives, with side-view recordings providing optimal results. AI-generated feedback achieved 4.08/5.0 expert quality ratings compared to 4.42/5.0 for human PGA professionals, with particular strength in comprehensiveness but weaker prioritization [2]. By combining objective biomechanical analysis with natural language coaching at a fraction of traditional instruction costs, Twelfth Tee represents a significant advancement in accessible sports training technology.

Golf Swing Analysis, Pose Estimation, AI Coaching, Sports Technology