titleRenderFormer-V2: Neural Rendering with Heterogeneous Scene Primitives
authorsChong Zeng, Yue Dong, Pieter Peers, Lvmin Zhang, Maneesh Agrawala
date2026-09-04
venuearXiv (preprint, not peer reviewed)
idarXiv:2609.05738
affiliationStanford AI
upvotes3 (Hugging Face Daily Papers, 2026-11-01)
linksabs pdf hf code
abstractWe present 'RenderFormer-V2', a unified learned transformer-based neural rendering model, complementary to modern physics-based rendering systems, that can handle diverse light-transport effects such as caustics, volumetric scattering, environment lighting, textured and displaced surfaces and out-of-distribution materials without per-scene training or specialized code. RenderFormer-V2 models global light transport as a sequence-to-sequence transformation. Following its predecessor, RenderFormer-V2 also employs a two stage process: a view-independent stage that resolves intra-scene primitive to primitive transport, and a view-dependent stage that transforms the internal neural scene representation into image pixels. Different from RenderFormer, our model employs a novel combined windowed-attention and rendering-informed attention sink in the view-independent stage to improve scalability while maintaining render accuracy. To further improve versatility, RenderFormerV2 supports heterogeneous scene primitives, including environment maps and participating media, and it employs a material encoding independent of the underlying surface reflectance model that encodes material appearance via a novel neural embedding. We demonstrate the versatility of RenderFormer-V2 on a variety of scenes and perform an extensive ablation of the improved attention mechanism.

cite

@misc{zeng2026renderformer,
  title={RenderFormer-V2: Neural Rendering with Heterogeneous Scene Primitives},
  author={Chong Zeng and Yue Dong and Pieter Peers and Lvmin Zhang and Maneesh Agrawala},
  year={2026},
  eprint={2609.05738},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2609.05738}
}