研究人员开发了F-RNG，一种新颖的前馈式框架，旨在从稀疏视图输入生成可重照明的3D高斯喷溅（3DGS）资产。该方法利用现有的大型重建模型（LRMs）和内在分解模型（IDMs）来提取可重照明的表示，而无需进行广泛的再训练。F-RNG通过外观的先验引导蒸馏来增强几何合成，能够实现灵活的重照明，与最先进的方法相比，计算成本显著降低，推理时间更快。 AI
影响 能够更快、更经济高效地创建可重照明的3D资产，可能对实时渲染和虚拟环境产生影响。
排序理由 该集群包含一篇详细介绍3D资产生成新方法的学术论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2605.25975v1 Announce Type: cross Abstract: Capturing relightable 3D assets from real-world objects is a widely researched problem. Several per-scene optimization-based methods, based on 3D Gaussian splatting (3DGS), support relighting; however, they usually require dense i…
Capturing relightable 3D assets from real-world objects is a widely researched problem. Several per-scene optimization-based methods, based on 3D Gaussian splatting (3DGS), support relighting; however, they usually require dense input views, and their overfitting nature makes it …