AI Paper Atlas

Research map / 3D vision / 3D generation

Future Dynamic 3D Reconstruction: A 3D World Model with Disentangled Ego-Motion

Nils Morbitzer, Jonathan Evers, Artem Savkin, Thomas Stauner, Nassir Navab, Federico Tombari, Stefano Gasperini

ICML 2026 · 3D vision › 3D generation · approach: Monocular 3D object reconstruction

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 information sharing between pose and spatial masked transformers via cross-attention (ablation)FR3D (A3 baseline) · Waymo 224x224, t+2.0s poseATE: 0.895 -> 0.389Improved
add information sharing between pose and spatial transformers (ablation)FR3D (A3 baseline) · Waymo 224x224, t+2.0s depthAbsRel: 0.183 -> 0.173Improved
add autoregressive training (ablation)FR3D (A3 baseline) · Waymo 224x224, t+2.0s depthAbsRel: 0.183 -> 0.179Improved
add autoregressive training (ablation)FR3D (A3 baseline) · Waymo 224x224, t+2.0s poseATE / RPE_R: ATE 0.895 -> 0.670, RPE_R 0.724 -> 0.988Mixed
combine autoregressive training and information sharing (ablation)FR3D · Waymo 224x224, t+2.0s depthAbsRel: 0.173 -> 0.158 (A5 -> A6)Improved
combine autoregressive training and information sharing (ablation)FR3D · Waymo 224x224, t+2.0s poseATE: 0.389 -> 0.442 (A5 -> A6)Mixed
add pose masked transformer to CUT3R-Foresight (ablation)CUT3R-Foresight · Waymo 224x224 depthAbsRel: 0.183 -> 0.183 at t+2.0sNo clear effect
fine-tune at higher resolution 512x336 instead of 224x224 (ablation)FR3D · Waymo depth, t+2.0sAbsRel: 0.158 -> 0.152Improved
use CUT3R latent tokens with disentangled pose forecasting instead of DINOv2 features (DINO-Foresight)future depth forecasting · nuScenes zero-shot, t+2.5sAbsRel: 0.283 -> 0.229Improved
disentangle ego-motion and forecast with FR3D instead of CUT3R-Foresight on static regionsfuture depth forecasting · Waymo static regions, t+2.0sAbsRel: 0.170 -> 0.145Improved
disentangle ego-motion and forecast with FR3D instead of CUT3R-Foresight on dynamic regionsfuture depth forecasting · Waymo dynamic regions, t+2.0sAbsRel: 0.353 -> 0.337Improved
forecast with FR3D rather than a 2D video world model (Vista)future prediction · 2s forecast, 4 context frameslatency (s): 24.57 -> 0.85Improved

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