AI Paper Atlas

Research map / 3D vision / 3D generation

SceneDirector: Bridging Explicit Geometry and Generative Priors for Unified Driving Scene Editing

Yiyuan Liang, Zhiying Yan, Tao Zhang, Shangke Liu, Kai Lin, Xu Zou, Nong Sang, Changxin Gao

ICML 2026 · 3D vision › 3D generation · approach: 3D perception for autonomous driving

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 Mask-Gated Reference Attention (reference injection) (ablation)SceneDirector · multi-view 3m deviationFVD: 495.0 -> 469.2Improved
add Mask-Gated Reference Attention (reference injection) (ablation)SceneDirector · multi-view 3m deviationFID: 53.02 -> 49.10Improved
add Mask-Gated Reference Attention (reference injection) (ablation)SceneDirector · multi-view 2m fixed offsetFVD: 512.5 -> 474.1Improved
add uncertainty-aware gating to cross-attention (MGRA vs standard attention) (ablation)SceneDirector · multi-view 3m deviationFVD: 481.8 -> 469.2Improved
add uncertainty-aware gating to cross-attention (MGRA vs standard attention) (ablation)SceneDirector · multi-view 2m fixed offsetFVD: 493.4 -> 474.1Improved
replace depth completion backbone with LRRU (ablation)SceneDirector · multi-view 2m fixed offsetFID: 51.34 -> 51.87No clear effect
replace depth completion backbone with LRRU (ablation)SceneDirector · multi-view 3m deviationFID: 49.10 -> 50.04Hurt
add object plus ego-trajectory editing jointly vs object-onlySceneDirector · multi-view object editingATE: 0.90 -> 0.95Hurt
unified single-pass editing vs sequential single-edit baselinesVACE · Multi-EditFVD: 729.42 -> 516.83Improved
use SceneDirector instead of VACE-14B for object insertionVACE · Single-EditATE: 1.02 -> 0.78Improved
use SceneDirector instead of VACE-14B for object insertionVACE · Single-EditFVD: 451.31 -> 464.56Hurt
train StreetGaussian on SceneDirector-generated videosStreetGaussian · front-view 2m gradual deviationFVD: 434.9 -> 400.1Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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