Behavioral Recognition and Classification of Thai Boxing Techniques Using Hybrid Deep Learning Models

Thai boxing (Muay Thai) is a combat sport gaining global popularity. This research applies hybrid deep learning to automatically detect and classify 10 out of 15 traditional Mae Mai Muay Thai behavioral patterns from videos. Using a curated dataset from a professional boxer, the model achieved 98% accuracy in identifying offensive/defensive maneuvers. Performance metrics (precision, recall, F1-score, and confusion matrix) validate its effectiveness for athlete analysis, coaching, and judging. Results indicate that hybrid deep learning can interpret Muay Thai’s dynamic movements, enhancing sports analytics.