Research map / 3D vision / 3D generation
MVCustom: Multi-View Customized Diffusion via Geometric Latent Rendering and Completion
Minjung Shin, Hyunin Cho, Sooyeon Go, Jin-Hwa Kim, Youngjung Uh
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use MVCustom instead of Txt-MV gen with DB | multi-view customization · CO3Dv2 multi-view customization | camera pose accuracy: 0.283 -> 0.735 | Improved |
| use MVCustom instead of CustomDiffusion360 | multi-view customization · CO3Dv2 multi-view customization | camera pose accuracy: 0.000 -> 0.735 | Improved |
| use MVCustom instead of Txt-MV gen with DB | multi-view customization · CO3Dv2 multi-view customization | multi-view consistency (lower better): 0.116 -> 0.121 | Hurt |
| use MVCustom instead of CustomDiffusion360 | multi-view customization · CO3Dv2 multi-view customization | text alignment: 0.806 -> 0.744 | Hurt |
| use MVCustom instead of CustomDiffusion360 | multi-view customization · CO3Dv2 multi-view customization | inference time and GPU memory: 74.97 -> 130.92 s; 4.99 -> 19.29 GB | Hurt |
| add depth-aware feature rendering to customization fine-tuning (ablation) | MVCustom · qualitative x-translation camera pose | geometric consistency of surroundings | Improved |
| add consistent-aware latent completion after feature rendering (ablation) | MVCustom · qualitative x-translation camera pose | realism of disoccluded regions | Improved |
| use dense spatio-temporal attention instead of 1D temporal attention (ablation) | MVCustom · feature replacement test (qualitative) | spatial consistency | Improved |
| use FeatureNeRF in more transformer blocks (ablation) | MVCustom · CO3Dv2 customization | identity preservation vs text responsiveness | Mixed |
Closest papers by meaning
- MVGenMaster: Scaling Multi-View Generation from Any Image via 3D Priors Enhanced Diffusion Model (CVPR 2025 · similarity 0.90)
- MVDream: Multi-view Diffusion for 3D Generation (ICLR 2024 · similarity 0.89)
- MEAT: Multiview Diffusion Model for Human Generation on Megapixels with Mesh Attention (CVPR 2025 · similarity 0.88)
- Zero-Shot Novel View and Depth Synthesis with Multi-View Geometric Diffusion (CVPR 2025 · similarity 0.87)
- NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis (NeurIPS 2026 · similarity 0.87)
- DreamComposer: Controllable 3D Object Generation via Multi-View Conditions (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
