Survey on enhancing athletic training with activity recognition and deep learning

The proposed methodology focuses on the advanced analysis and enhancement of athletic techniques, particularly targeting the precision, angle, and positioning of movements such as shooting. By integrating technologies like Human Action Recognition (HAR), Artificial Intelligence (AI), and Deep Learning, the system analyzes player data from images or videos, identifying errors and offering insights for improvement. It suggests optimal shooting angles, positions for maximum scoring, and corrective measures to refine technique, thus aiding coaches in player assessment, strategy formulation, and tactical decision-making. The methodology employs OpenCV and Machine Learning algorithms for accurate performance analysis, while Deep Learning models, such as Artificial Neural Networks (ANN) and You Only Look Once (YOLOv8), optimize feature extraction and analysis. YOLOv8, an advanced computer vision framework, ensures precise detection of key attributes. These combined technologies enable the identification of performance flaws and guide athletes toward achieving their goals. The solution is developed using Python, OpenCV, HAR, and YOLOv8, with IDEs like VSCode and Jupyter Notebook facilitating its implementation.