AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

LaneDiffusion: Improving Centerline Graph Learning via Prior Injected BEV Feature Generation

Zijie Wang, Weiming Zhang, Wei Zhang, Xiao Tan, Hongxing Liu, Yaowei Wang, Guanbin Li

ICCV 2025 · Graph learning › Graph self-supervised learning · approach: Diffusion-based graph generation · 3 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 LaneDiffusion BEV prior diffusion to CGNetCGNet · nuScenes point-levelGEO F1: 54.7 -> 58.9Improved
add LaneDiffusion BEV prior diffusion to CGNetCGNet · nuScenes point-levelTOPO F1: 42.2 -> 46.8Improved
add LaneDiffusion BEV prior diffusion to CGNetCGNet · nuScenes point-levelAPLS: 30.7 -> 37.1Improved
add LaneDiffusion BEV prior diffusion to CGNetCGNet · nuScenes segment-levelmAP_cf: 35.2 -> 41.6Improved
add LaneDiffusion BEV prior diffusion to CGNetCGNet · nuScenes segment-levelTOP_ll: 1.3 -> 3.4Improved
add LaneDiffusion BEV prior diffusion to CGNetCGNet · Argoverse2 valTOPO F1: 44.5 -> 47.0Improved
add LaneDiffusion BEV prior diffusion to CGNetCGNet · Argoverse2 valTOP_ll: 0.5 -> 1.8Improved
use T=15 sampling steps instead of T=5 (ablation)LaneDiffusion · nuScenesAPLS: 33.1 -> 37.1Improved
use T=30 sampling steps instead of T=15 (ablation)LaneDiffusion · nuScenesTOPO F1: 46.8 -> 46.0Hurt
train LPDM with Overall paradigm instead of Sampling (T=5) (ablation)LaneDiffusion · nuScenesTOPO F1: 44.8 -> 45.6Improved
add feature refinement (Concat&ED) after generation (ablation)LaneDiffusion · nuScenesTOPO F1: 43.9 -> 46.8Improved
use encoder-decoder instead of 1x1 conv for refinement (concat) (ablation)LaneDiffusion · nuScenesAPLS: 35.7 -> 37.1Improved

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