Development of an Age and Gender Prediction System in Sports using Deep Learning

Abstract

Age and gender are critical in sports but often lead to abuse and unfair practices. This study develops a deep learning system to address these issues by improving the accuracy of age and gender prediction. Traditional methods frequently mismatch athletes' actual capabilities with their competition categories. The proposed system, using advanced neural networks and diverse datasets, aims to provide precise physiological age and gender predictions, aiding coaches, sports organizations, and medical professionals. The research includes a literature review on existing models and deep learning techniques, and the methodology involves data preprocessing, model development, and performance evaluation. The system achieved an 82% success rate in gender prediction and 69% in age prediction, showing promise for enhancing fairness in sports competitions. However, accuracy and fairness limitations emphasize the need for ongoing refinement.