AI Paper Atlas

Research map / Video / Video generation

WoVoGen: World Volume-aware Diffusion for Controllable Multi-camera Driving Scene Generation

Jiachen Lu, Ze Huang, Zeyu Yang, Zhang Jiahui, Li Zhang

ECCV 2024 · Video › Video generation · approach: World-model video generation · 62 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
condition multi-view diffusion on a dense world volume rather than sparse conditionsmulti-camera street-view generation · nuScenes validationFID: 52.6 -> 27.6 (vs DriveDreamer)Improved
generate multi-view video with temporal finetuning and world volumeWoVoGen · nuScenes validationFVD: 452.0 -> 417.7 (vs DriveDreamer)Improved
add world volume condition to object guidance (ablation)WoVoGen single-frame · nuScenes validationFID: 28.6 -> 27.6Improved
add object guidance to world volume (ablation)WoVoGen single-frame · nuScenes validationFID: 27.8 -> 27.6Improved
add temporal attention finetuning on video generation (ablation)WoVoGen · qualitative front/back camera sequencestemporal consistencyImproved
add generated data to the nuScenes training set for BEVDetBEVDet · nuScenes 3D detection (cars, trucks, buses)mAP_v: 34.9 -> 36.2Improved
add generated data to the training setBEVDet · nuScenes 3D detectionmAOE_v: 17.9 -> 15.7Improved

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