AI Paper Atlas

Research map / 3D vision / 3D generation

La La LiDAR: Large-Scale Layout Generation from LiDAR Data

Youquan Liu, Lingdong Kong, Weidong Yang, Xin Li, Alan Liang, Runnan Chen, Ben Fei, Tongliang Liu

AAAI 2026 · 3D vision › 3D generation · approach: Diffusion-based novel view synthesis · 11 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 semantic graph features to layout generation (ablation)La La LiDAR layout · nuScenes layout generation ablationRAD: 0.59 -> 0.68Improved
add semantic graph features to layout generation (ablation)La La LiDAR layout · nuScenes layout generation ablationIoU (%): 23.64 -> 28.14Improved
use cross-attention conditioning instead of concatenation in layout diffusion (ablation)La La LiDAR layout · nuScenes layout generation ablationRAD: 0.63 -> 0.68Improved
use cross-attention conditioning instead of concatenation in layout diffusion (ablation)La La LiDAR layout · nuScenes layout generation ablationPrec0.5 (%): 4.82 -> 6.15Improved
add collision and IoU losses to layout training (ablation)La La LiDAR layout · nuScenes layout generation ablationcollision rate: 0.08 -> 0.06Improved
add foreground validity mask to scene generation (ablation)La La LiDAR scene · nuScenes scene generation ablationFPD: 19.5 -> 9.8Improved
use scene graph conditioning instead of box layout condition (ablation)La La LiDAR scene · nuScenes scene generation ablationFRD: 244.6 -> 211.0Improved
use scene graph conditioning instead of box layout condition (ablation)La La LiDAR scene · nuScenes scene generation ablationFPD: 14.2 -> 9.8Improved
use FCI instead of cross-attention-only conditioning (ablation)La La LiDAR scene · nuScenes scene generation ablationFPD: 27.5 -> 9.8Improved
use FCI instead of addition or concatenation conditioning (ablation)La La LiDAR scene · nuScenes scene generation ablationFRD: 235.37 -> 211.03Improved
use La La LiDAR instead of R2DM for generative data augmentationMinkUNet · nuScenes semantic segmentation val (1% labels)mIoU: 64.1 -> 65.1Improved
use La La LiDAR instead of LaserMix at 10% labelsSPVCNN · nuScenes semantic segmentation val (10% labels)mIoU: 74.1 -> 74.0Hurt

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