AI Paper Atlas

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

ICLR 2026 · 3D vision › 3D generation · approach: Diffusion-based novel view synthesis · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
use MVCustom instead of Txt-MV gen with DBmulti-view customization · CO3Dv2 multi-view customizationcamera pose accuracy: 0.283 -> 0.735Improved
use MVCustom instead of CustomDiffusion360multi-view customization · CO3Dv2 multi-view customizationcamera pose accuracy: 0.000 -> 0.735Improved
use MVCustom instead of Txt-MV gen with DBmulti-view customization · CO3Dv2 multi-view customizationmulti-view consistency (lower better): 0.116 -> 0.121Hurt
use MVCustom instead of CustomDiffusion360multi-view customization · CO3Dv2 multi-view customizationtext alignment: 0.806 -> 0.744Hurt
use MVCustom instead of CustomDiffusion360multi-view customization · CO3Dv2 multi-view customizationinference time and GPU memory: 74.97 -> 130.92 s; 4.99 -> 19.29 GBHurt
add depth-aware feature rendering to customization fine-tuning (ablation)MVCustom · qualitative x-translation camera posegeometric consistency of surroundingsImproved
add consistent-aware latent completion after feature rendering (ablation)MVCustom · qualitative x-translation camera poserealism of disoccluded regionsImproved
use dense spatio-temporal attention instead of 1D temporal attention (ablation)MVCustom · feature replacement test (qualitative)spatial consistencyImproved
use FeatureNeRF in more transformer blocks (ablation)MVCustom · CO3Dv2 customizationidentity preservation vs text responsivenessMixed

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy