Research map / Detection and segmentation / 3D detection and driving
Towards Accurate 3D Object Detection in Adverse Weather by Leveraging 4D Radar for LiDAR Geometry Enhancement
Tianxu Tong, Xinrun Liu, Hongmin Liu, Bin Fan
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add Virtual LiDAR Generation to the fusion baseline (ablation) | fusion baseline · K-Radar V1.0 (Sedan, all weather) | AP3D: 46.0 -> 50.4 | Improved |
| add Position-Guided Cross-Attention to baseline with VLG (ablation) | REL · K-Radar V1.0 (Sedan, all weather) | AP3D: 50.4 -> 52.5 | Improved |
| add Adaptive Feature Fusion to baseline with VLG and PGCA (ablation) | REL · K-Radar V1.0 (Sedan, all weather) | AP3D: 52.5 -> 54.5 | Improved |
| add all REL modules (VLG, PGCA, AFF) to the baseline (ablation) | fusion baseline · K-Radar V1.0 (Sedan, all weather) | FPS: 24.1 -> 20.7 | Hurt |
| deploy modules in stages II and III instead of only stage II (ablation) | REL · K-Radar V1.0 (Sedan, all weather) | AP3D: 51.6 -> 54.5 | Improved |
| deploy modules in stage I only instead of II+III (ablation) | REL · K-Radar V1.0 (Sedan, all weather) | AP3D: 54.5 -> 49.5 | Hurt |
| deploy modules in all three stages instead of II+III (ablation) | REL · K-Radar V1.0 (Sedan, all weather) | AP3D: 54.5 -> 51.3 | Hurt |
| integrate REL into L4DR fusion framework | L4DR · K-Radar V2.0 Bus or Truck (adverse weather) | AP3D: 43.9 -> 48.7 | Improved |
| integrate REL into L4DR fusion framework | L4DR · K-Radar V2.0 Sedan (adverse weather) | AP3D: 54.0 -> 54.5 | Improved |
| replace 3D-LRF fusion with REL | 3D-LRF · K-Radar V1.0 (Sedan, all weather) | AP3D: 45.2 -> 54.5 | Improved |
| replace 3D-LRF fusion with REL | 3D-LRF · Vod-Fog level 4 Car (Moderate) | 3D mAP: 20.6 -> 49.7 | Improved |
| replace 3D-LRF fusion with REL | 3D-LRF · K-Radar V1.0 (Normal) | AP3D: 45.3 -> 53.0 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- L4DR — L4DR: LiDAR-4DRadar Fusion for Weather-Robust 3D Object Detection (AAAI 2025) · better on 5 of 5 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
- L4DR: LiDAR-4DRadar Fusion for Weather-Robust 3D Object Detection (AAAI 2025 · similarity 0.93)
- Towards Robust 3D Object Detection with LiDAR and 4D Radar Fusion in Various Weather Conditions (CVPR 2024 · similarity 0.92)
- An Empirical Study of the Generalization Ability of Lidar 3D Object Detectors to Unseen Domains (CVPR 2024 · similarity 0.87)
- Sunshine to Rainstorm: Cross-Weather Knowledge Distillation for Robust 3D Object Detection (AAAI 2024 · similarity 0.87)
- Equirectangular Point Reconstruction for Domain Adaptive Multimodal 3D Object Detection in Adverse Weather Conditions (AAAI 2025 · similarity 0.87)
- TransLoc4D: Transformer-based 4D Radar Place Recognition (CVPR 2024 · similarity 0.86)
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.
