AI Paper Atlas

Research map / Point clouds / Segmentation and detection

3D Object Detection with a Self-Supervised Lidar Scene Flow Backbone

Emeç Erçelik, Ekim Yurtsever, Mingyu Liu, Zhijie Yang, Hanzhen Zhang, Pınar Topçam, Maximilian Listl, Yılmaz Kaan Çaylı, Alois Knoll

ECCV 2022 · Point clouds › Segmentation and detection · approach: LiDAR voxel and pillar detectors · 30 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
initialize backbone with self-supervised scene flow pre-trainingPoint-GNN · KITTI val car3D AP Hard: 77.70 -> 80.12Improved
initialize backbone with self-supervised scene flow pre-trainingPointPillars · KITTI val car3D AP Hard: 67.76 -> 74.32Improved
initialize backbone with self-supervised scene flow pre-trainingPointPillars · KITTI test car3D AP Mod: 68.57 -> 72.99Improved
initialize backbone with self-supervised scene flow pre-trainingPointPillars · KITTI test carBEV AP Easy: 90.74 -> 88.92Hurt
initialize backbone with self-supervised scene flow pre-trainingPointPillars · nuScenes valmAP: 40.02 -> 42.06Improved
initialize backbone with self-supervised scene flow pre-trainingCenterPoint · nuScenes valmAP: 49.13 -> 49.94Improved
initialize backbone with self-supervised scene flow pre-trainingCenterPoint · nuScenes testmAP: 49.54 -> 51.42Improved
pre-train with scene flow when only 1% labeled data used (ablation)Point-GNN · KITTI val car (1% labels)3D AP Easy: 63.34 -> 66.47Improved
pre-train with scene flow when 20% labeled data used (ablation)Point-GNN · KITTI val car (20% labels)3D AP Easy: 88.47 -> 88.65No clear effect
add alternating training between flow and detection (ablation)Point-GNN · KITTI val car3D AP Hard: 2.32%Improved
use scene flow pre-training instead of GCC3DCenterPoint · nuScenes val (5% labels)mAP: 32.75 -> 36.04Improved
use scene flow pre-training instead of PointContrastCenterPoint · nuScenes val (10% labels)mAP: 38.25 -> 41.29Improved

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