The world’s first prototype of an automated solution for attack detection in semi-contact sports was presented – specifically for Karate Kumite (sparring). This AI-powered technology analyzes video footage using machine learning and advanced image processing to revolutionize competition scoring.
Technical Implementation
The technical implementation uses YOLOv3 for participant identification and LSTM networks for analyzing attack sequences. Tests showed a 97.47% classification accuracy for technique recognition and point scoring.
The system detects potential rule violations with a precision previously unavailable to referees.
Edvin Kuric developed this innovation in 2024, which represents a decisive step for modern karate and demonstrates how modern technology promotes fairness and precision.