Researchers have developed F-RNG, a novel feed-forward framework designed to generate relightable 3D Gaussian Splatting (3DGS) assets from sparse-view inputs. This approach leverages existing large reconstruction models (LRMs) and intrinsic decomposition models (IDMs) to extract relightable representations without requiring extensive retraining. F-RNG enhances geometry synthesis and uses prior-guided distillation for appearance, enabling flexible relighting with significantly reduced computational cost and faster inference times compared to state-of-the-art methods. AI
IMPACT Enables faster and more cost-effective creation of relightable 3D assets, potentially impacting real-time rendering and virtual environments.
RANK_REASON The cluster contains a research paper detailing a new method for 3D asset generation.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
