Research map / Detection and segmentation / 3D detection and driving
Rethinking IoU-Based Optimization for Single-Stage 3D Object Detection
Hualian Sheng, Sijia Cai, Na Zhao, Bing Deng, Jianqiang Huang, Xian-Sheng Hua, Min-Jian Zhao, Gim Hee Lee
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 RDIoU losses (ablation) | PointPillar · KITTI val Car Moderate | 3D R40 AP: 79.88 -> 82.42 | Improved |
| add RDIoU losses (ablation) | SECOND · KITTI val Car Moderate | 3D R40 AP: 82.85 -> 85.80 | Improved |
| add RDIoU losses (ablation) | CT-stacked · KITTI val Car Moderate | 3D R40 AP: 83.01 -> 86.20 | Improved |
| use RDIoU-guided instead of 3D IoU-guided IoU loss (ablation) | CT-stacked · KITTI val Car Moderate | 3D R40 AP: 83.60 -> 85.73 | Improved |
| use RDIoU-guided instead of 3D IoU-guided DIoU loss (ablation) | CT-stacked · KITTI val Car Moderate | 3D R40 AP: 84.20 -> 86.20 | Improved |
| use RDIoU-guided instead of 3D IoU-guided CIoU loss (ablation) | CT-stacked · KITTI val Car Moderate | 3D R40 AP: 84.01 -> 85.98 | Improved |
| replace Smooth-l1 with 3D IoU-guided DIoU loss (ablation) | CT-stacked · KITTI val Car Moderate | 3D R40 AP: 83.14 -> 84.20 | Improved |
| replace traditional focal loss with 3D IoU-guided QFL (ablation) | CT-stacked · KITTI val Car Moderate | 3D R40 AP: 85.12 -> 85.79 | Improved |
| replace 3D IoU-guided QFL with RDIoU-guided QFL (ablation) | CT-stacked · KITTI val Car Moderate | 3D R40 AP: 85.79 -> 86.20 | Improved |
| set rotation edge k to 1.0 rather than 0.6 (ablation) | RDIoU · KITTI val Car Moderate | 3D R40 AP: 85.89 -> 86.20 | Improved |
| set k larger than 1.0 (1.8) (ablation) | RDIoU · KITTI val Car Moderate | 3D R40 AP: 86.20 -> 85.59 | Hurt |
| use RDIoU instead of ODIoU (ablation) | SECOND · KITTI val Car Moderate | 3D R40 AP: 83.88 -> 85.80 | 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.
- Voxel-RCNN — Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detection (AAAI 2021) · better on 1 of 1 shared result
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
- Deep Dive Into Gradients: Better Optimization for 3D Object Detection With Gradient-Corrected IoU Supervision (CVPR 2023 · similarity 0.93)
- Generalized Intersection Over Union: A Metric and a Loss for Bounding Box Regression (CVPR 2019 · similarity 0.89)
- Rotation-robust Intersection over Union for 3D Object Detection (ECCV 2020 · similarity 0.88)
- Rethinking Boundary Discontinuity Problem for Oriented Object Detection (CVPR 2024 · similarity 0.88)
- 3DIoUMatch: Leveraging IoU Prediction for Semi-Supervised 3D Object Detection (CVPR 2021 · similarity 0.87)
- Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression (AAAI 2020 · 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.
