PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation
State-of-the-art single-image 3D reconstruction methods often rely on complex hybrid architectures and loss functions, or compress geometry into latent spaces in order to leverage pre-trained latent diffusion models. In this work, we show that such architectural overhead and intricate loss formulations are unnecessary. We introduce a minimalist pixel-space Diffusion Transformer, built on a plain ViT, that operates directly on raw 3D point map patches and is conditioned on image tokens from a pre-trained DINOv3. Unlike existing latent diffusion approaches, we train our diffusion backbone entire
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 46%Diffuse-XL →
- LinkedLinked via arxiv author · 85%Haofei Xu →
“PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation”
- LinkedLinked via arxiv author · 85%Rundi Wu →
“PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation”
- LinkedLinked via arxiv author · 85%Philipp Henzler →
“PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation”
- LinkedLinked via arxiv author · 85%Nikolai Kalischek →
“PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation”
- LinkedLinked via arxiv author · 85%Michael Oechsle →
“PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation”
- LinkedLinked via arxiv author · 85%Fabian Manhardt →
“PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation”
- LinkedLinked via arxiv author · 85%Marc Pollefeys →
“PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation”