Paper Detail
Mayank Singh, Michele Stoppa, Alvise Memo, Rui Yu, Harsha Kalli, Srimanth Gunturi, Muhammad Ahmed Riaz, Behrooz Shahsavari, Waleed Abdulla, David E. Jacobs
High-fidelity image-to-3D generation requires a 3D representation that captures both geometry and appearance. To support relighting and integration into standard rendering pipelines, the representation should include physically based rendering (PBR) modalities such as albedo, metallic-roughness, and surface normals. We propose Luce, a 3D representation that unifies geometry and PBR materials within a voxelized multimodal Gaussian cloud, using dedicated Gaussian primitives for each modality. A variational autoencoder compresses this representation into a unified material-aware latent space. A rectified-flow transformer generates this latent from a single image, conditioned on multi-layer features from a pretrained image encoder that preserve both semantic context and fine spatial detail. The latent then decodes into relightable PBR Gaussians and an optional textured mesh with a tangent-space normal map. On Toys4K, Luce achieves state-of-the-art single-image-to-3D generation, improving FID by 28% over the strongest baseline. We further introduce a benchmark of AI-generated images, on which Luce improves the CLIP image-alignment score over the best baseline (0.8519 vs. 0.8299). Luce generates relightable, geometrically accurate, and materially faithful assets that preserve fine details such as text, logos, and inscriptions.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@misc{singh2026luce,
title = {Luce: Relightable Gaussians for 3D Asset Generation},
author = {Mayank Singh and Michele Stoppa and Alvise Memo and Rui Yu and Harsha Kalli and Srimanth Gunturi and Muhammad Ahmed Riaz and Behrooz Shahsavari and Waleed Abdulla and David E. Jacobs},
year = {2026},
abstract = {High-fidelity image-to-3D generation requires a 3D representation that captures both geometry and appearance. To support relighting and integration into standard rendering pipelines, the representation should include physically based rendering (PBR) modalities such as albedo, metallic-roughness, and surface normals. We propose Luce, a 3D representation that unifies geometry and PBR materials within a voxelized multimodal Gaussian cloud, using dedicated Gaussian primitives for each modality. A va},
url = {https://huggingface.co/papers/2608.23943},
keywords = {voxelized multimodal Gaussian cloud, PBR materials, variational autoencoder, rectified-flow transformer, multi-layer features, tangent-space normal map, huggingface daily},
eprint = {2608.23943},
archiveprefix = {arXiv},
}
{}