High Fidelity Rendering
High-fidelity rendering aims to create photorealistic and geometrically accurate 3D representations from various input sources, such as images or videos. Current research heavily focuses on improving existing neural rendering methods, particularly those based on neural radiance fields (NeRFs) and Gaussian splatting, by addressing limitations in handling complex geometries, dynamic scenes, and uncertainty. These advancements are significant for applications in virtual and augmented reality, computer graphics, and 3D modeling, enabling more immersive and realistic experiences.
Papers
NPBG++: Accelerating Neural Point-Based Graphics
Ruslan Rakhimov, Andrei-Timotei Ardelean, Victor Lempitsky, Evgeny Burnaev
Learning Motion-Dependent Appearance for High-Fidelity Rendering of Dynamic Humans from a Single Camera
Jae Shin Yoon, Duygu Ceylan, Tuanfeng Y. Wang, Jingwan Lu, Jimei Yang, Zhixin Shu, Hyun Soo Park