Sports Highlights Generation - Master Thesis

You will work with advanced AI approaches to automate events detection and highlights extraction from indoor Sports game-play. The basis for the work is time synchronized multi-camera video streams overlooking the sports venue from several angles, thus already has the core ability to observe and track players and ball.
Your challenge will be to research and prototype methods for automatically deriving saliency and relevancy of essential game-play action (e.g. scoring, steals, and other events), perform camera selection and assemble watchable sports highlights.
The project goal is to build and evaluate a prototype pipeline performing these tasks using recent-most know approaches.
Who are you:
We expect you are 1 or 2 students, with basic knowledge or hands-on experience of Visual Language Models (VLM), Transformers, etc.
About You:
- A Master’s degree student in computer science, mathematical science, electrical engineering, or similar.
- Have a strong interest in c...