AI Paper Atlas

Research map / Detection and segmentation / 3D detection and driving

Benchmarking Robustness of 3D Object Detection to Common Corruptions

Yinpeng Dong, Caixin Kang, Jinlai Zhang, Zijian Zhu, Yikai Wang, Xiao Yang, Hang Su, Xingxing Wei, Jun Zhu

CVPR 2023 · Detection and segmentation › 3D detection and driving · approach: LiDAR perception for 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
Use LiDAR-camera fusion rather than camera-only detection3D object detectors · nuScenes-CmAPcor: 10.26; 10.68; 18.60; 22.79 -> 56.99; 58.73; 61.03Improved
Use LiDAR-camera fusion rather than LiDAR-only detection3D object detectors · nuScenes-CmAPcor: 23.42; 40.37; 49.81 -> 56.99; 58.73; 61.03Improved
Use LiDAR-camera fusion rather than camera-only detection3D object detectors · KITTI-C, car moderateAPcor: 2.68; 2.42; 3.05 -> 67.81; 71.87Improved
Apply motion-level corruptions3D object detectors · KITTI-CAPHurt
Apply motion-level corruptionsLiDAR-only and LiDAR-camera fusion detectors · nuScenes-CmAPHurt
Apply weather-level corruptions3D detectors trained on normal-weather data · KITTI-CAPHurt
Apply weather-level corruptionsnuScenes-C detectors · nuScenes-CmAPNo clear effect
Corrupt both LiDAR and camera inputs rather than one modalityFocals Conv · KITTI-CAPHurt
Use multi-view detection rather than monocular detectionCamera-only detectors · nuScenes-CmAP: 57%; 54% -> 46%; 45%Improved
Use BEVFormer rather than DETR3DCamera-only detectors · nuScenes-C, object-level corruptionsmAPImproved
Use LiDAR-only rather than fusion detection for corruptions affecting both modalities3D object detectors · KITTI-CAPMixed
Use fusion detection rather than LiDAR-only detection for point-cloud-only corruptions3D object detectors · KITTI-CAPImproved

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