AI-Based Performance Analysis in TV Sports Programs

In this study, we examine the use of artificial intelligence (AI) technologies in televised sports programmers to facilitate real-time performance analysis and improve audience engagement. Sport broadcasting is mainly dependent on the way of providing manual comments and post-game analyses, where results are frequently without any immediacy as well as impartiality. To ameliorate this project proposes a unique multimodal AI driven methodology that utilizes computer vision, natural language generation, and audio-visual sync to dissect key performance metrics such as player movements, tempo, and emotional signals of players. “Leveraging over 150 hours of annotated datasets across football, basketball, and tennis, this study presents a lowerror convolutional neural networks to detect events, track players from sports broadcasts and investigate the sentiment of commentary using attention-based bidirectional LSTM models. The experimental outcomes show high accuracy for event detection with 96 % for goals, as well as strong contextual agreement between AI-generated insights and commentary provided by humans. The findings highlight how AI can help automate and enhance the live sports viewing experience, providing another leap in intelligent sports media production.