• Topic: Implementing and analysing a deep learning framework that enables arbitrary-scale super-resolution for real-time rendering using Fourier-based implicit neural representations. This method allows rendering frames super-resolved to any scale with a single trained model, providing high-quality results in real-time with minimal artefacts.
• Year: 2024
• YouTube Presentation: https://www.youtube.com/watch?v=AkegnIRF_S4
• Students: Gad Azriel & Adar Budomski