DEVELOPMENT OF AN INTELLIGENT VIDEO ANALYTICS MODEL FOR EVALUATING WRESTLING TECHNIQUES

Dehkanov Abdulaziz Ilhomjon ugli

Authors

Dehkanov Abdulaziz Ilhomjon ugli Lecturer at the Department of Theory and Methodology of Martial Arts, Fergana State University

Keywords:

wrestling, artificial intelligence, video analytics, computer vision, technique evaluation, biomechanics, deep learning.

Abstract

This study presents the development of an intelligent video analytics model designed to evaluate technical performance in wrestling using artificial intelligence (AI) and computer vision algorithms. The proposed system automatically detects, classifies, and assesses wrestling techniques in real time, based on kinematic and biomechanical parameters. A dataset of 2,000 annotated video sequences from freestyle and Greco-Roman wrestling competitions was used to train the model. Using convolutional neural networks (CNN) and pose estimation frameworks (OpenPose, Mediapipe), the model achieved a recognition accuracy of 92.8% and an error rate below 0.15 s in detecting technical actions. The implementation of this system enables coaches to receive objective feedback, quantify technique efficiency, and enhance training personalization through data-driven analysis.

References

Baca, A., & Kornfeind, P. (2021). Computer Vision and Artificial Intelligence in Sports Performance Analysis. Springer.

Bompa, T., & Haff, G. (2019). Periodization: Theory and Methodology of Training. Human Kinetics.

Cao, Z., Hidalgo, G., Simon, T., Wei, S. E., & Sheikh, Y. (2019). “OpenPose: Realtime Multi-Person 2D Pose Estimation.” IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(1), 172–186.

Mooney, R., et al. (2020). “Deep Learning Approaches in Sports Motion Analysis.” Sports Engineering, 23(1), 9–18.

Petrov, V., & Abdurakhmanov, A. (2022). “Digital Video Analysis in Wrestling Training.” International Journal of Combat Sports Science, 4(2), 33–42.

Tang, Z., et al. (2021). “Hybrid CNN-LSTM Models for Human Action Recognition.” Pattern Recognition Letters, 146, 67–74.

United World Wrestling (2023). Digital Analytics and Performance Evaluation Standards. Lausanne: UWW Press.