Temporal Supersampling
Temporal supersampling aims to enhance frame rates for high-refresh-rate displays by intelligently predicting intermediate frames from existing rendered frames, mitigating visual artifacts like screen tearing and stuttering. Current research focuses on developing efficient extrapolation methods, often employing neural networks or hybrid approaches combining neural networks with faster warping techniques, to achieve high-quality results with minimal latency. These advancements are crucial for real-time applications like gaming and virtual reality, improving visual fidelity and user experience in high-frequency display environments.
Papers
NeRRF: 3D Reconstruction and View Synthesis for Transparent and Specular Objects with Neural Refractive-Reflective Fields
Xiaoxue Chen, Junchen Liu, Hao Zhao, Guyue Zhou, Ya-Qin Zhang
AxOCS: Scaling FPGA-based Approximate Operators using Configuration Supersampling
Siva Satyendra Sahoo, Salim Ullah, Soumyo Bhattacharjee, Akash Kumar