Smart Sports Studio: AI-Powered Tools for Next-Gen Media Production

This PhD project is set to transform the way sports media is produced by creating an innovative AI system that makes the process faster, cheaper, and more efficient. Imagine a system that can automatically detect key moments in a game, generate highlight reels, and seamlessly integrate multimedia content—all with minimal human intervention. This is exactly what this project aims to achieve by developing a modular AI framework that can be customized to fit the needs of various sports organizations, from large media companies to smaller clubs. The heart of this system is a conversational AI agent that makes it easy for users to get the results they want without needing to understand complex technology. Users simply tell the system what they need, and the AI takes care of the rest—choosing the right tools and combining them in the best way to get the job done. This could include anything from recognizing important events in a game, cutting video clips, adjusting the video format, to summarizing what happened in both text and video. By automating these time-consuming and expensive tasks, this AI system will not only save money and resources but also open up new possibilities for creativity and innovation in sports media. It will allow content creators to work more efficiently, delivering exciting and customized content that meets the diverse needs of audiences around the world.

This PhD project aims to revolutionize sports media production by developing a modular, multimodal AI framework that significantly reduces labor, cost, and time when extracting content and creating highlight summaries. As the industry increasingly demands more efficient, low cost and adaptive technologies, this plug-and-play system integrates various AI modules to automate complex tasks such as event detection, highlight generation, multimedia integration, and other tedious and labor-intensive tasks, which are particularly costly for growing sports industries, media, leagues and clubs. Central to this innovation is a conversational AI agent that simplifies user interaction, enabling easy customization and efficient operation tailored to specific project requirements. Users of this modular system only need to specify the task they want to accomplish; the conversational agent then selects, optimizes, and integrates the necessary developed internal AI modules — such as basic models for object detection, segmentation, tracking, and audio analyses, and then combined into more complex models for semantic understanding, event extraction, video clipping, aspect ration cropping, video and text summarisation, etc. The framework also includes intelligent algorithms for module configuration that optimize performance and resource utilization based on user input. This comprehensive approach not only sets new standards for efficiency and innovation in sports media production, but also enhances the ability of content creators to meet diverse production needs and viewer preferences with unprecedented agility and creativity.