Timezone: »

NerfDiff: Single-image View Synthesis with NeRF-guided Distillation from 3D-aware Diffusion
Jiatao Gu · Alex Trevithick · Kai-En Lin · Joshua M Susskind · Christian Theobalt · Lingjie Liu · Ravi Ramamoorthi

Tue Jul 25 02:00 PM -- 04:30 PM (PDT) @ Exhibit Hall 1 #231

Novel view synthesis from a single image requires inferring occluded regions of objects and scenes whilst simultaneously maintaining semantic and physical consistency with the input. Existing approaches condition neural radiance fields (NeRF) on local image features, projecting points to the input image plane, and aggregating 2D features to perform volume rendering. However, under severe occlusion, this projection fails to resolve uncertainty, resulting in blurry renderings that lack details. In this work, we propose NerfDiff, which addresses this issue by distilling the knowledge of a 3D-aware conditional diffusion model (CDM) into NeRF through synthesizing and refining a set of virtual views at test-time. We further propose a novel NeRF-guided distillation algorithm that simultaneously generates 3D consistent virtual views from the CDM samples, and finetunes the NeRF based on the improved virtual views. Our approach significantly outperforms existing NeRF-based and geometry-free approaches on challenging datasets including ShapeNet, ABO, and Clevr3D.

Author Information

Jiatao Gu (Apple (MLR))
Alex Trevithick (UC San Diego)
Kai-En Lin (University of California, San Diego)
Joshua M Susskind (Apple, Inc.)
Christian Theobalt (Max-Planck-Institute for Informatics, Saarland Informatics Campus)
Lingjie Liu (University of Pennsylvania, University of Pennsylvania)
Ravi Ramamoorthi (University of California, San Diego)

More from the Same Authors