AI Paper Atlas

Research map / 3D vision / 3D generation

Aligned Novel View Image and Geometry Synthesis via Cross-modal Attention Instillation

Min-Seop Kwak, Junho Kim, Sangdoo Yun, Dongyoon Han, Taekyung Kim, Seungryong Kim, Jin-Hwa Kim

ICLR 2026 · 3D vision › 3D generation · approach: Diffusion-based novel view synthesis · 4 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
add pointmap conditioning to baseline (ablation)image diffusion NVS · RealEstate10K extrapolativePSNR: 16.55 -> 16.93Improved
add proximity-based mesh conditioning (ablation)pointmap-conditioned model · RealEstate10K extrapolativePSNR: 16.93 -> 17.01Improved
add cross-modal attention instillation (ablation)full model · RealEstate10K extrapolativePSNR: 17.01 -> 17.41Improved
add all components to baseline (ablation)image diffusion NVS · RealEstate10K extrapolativeLPIPS: 0.260 -> 0.229Improved
increase reference views from 2 to 3 at inference (trained on 2)Ours · RealEstate10KPSNR: 17.41 -> 20.02Improved
increase reference views from 3 to 4Ours · RealEstate10KPSNR: 20.02 -> 20.08No clear effect
increase reference views from 1 to 7Ours · Co3DPSNR: 14.62 -> 23.45Improved
add Gaussian perturbation sigma 15% to pointmapOurs · DTU extrapolative 2-viewPSNR: 15.580 -> 13.726Hurt
mask 80% of predicted pointmap pointsOurs · DTU extrapolative 2-viewPSNR: 15.580 -> 13.000Hurt
mask 30% of predicted pointmap pointsOurs · DTU extrapolative 2-viewLPIPS: 0.184 -> 0.223Hurt
use Ours instead of NoPoSplat in interpolative settingOurs vs NoPoSplat · RealEstate10K interpolativePSNR: 25.03 -> 24.23Hurt
use Ours instead of LVSM (2-view)Ours · DTU extrapolativePSNR: 15.23 -> 15.58Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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