Research map / Detection and segmentation / 3D detection and driving
Cross-Modality Knowledge Distillation Network for Monocular 3D Object Detection
Yu Hong, Hang Dai, Yong Ding
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 depth pre-training of the backbone (ablation) | CMKD · KITTI val Car Moderate | 3D AP: 8.52 -> 13.48 | Improved |
| add response distillation (ablation) | CMKD · KITTI val Car Moderate | 3D AP: 13.48 -> 14.49 | Improved |
| add feature distillation (ablation) | CMKD · KITTI val Car Moderate | 3D AP: 13.48 -> 15.46 | Improved |
| add both distillation modules (ablation) | CMKD · KITTI val Car Moderate | 3D AP: 13.48 -> 16.33 | Improved |
| add unlabeled data distillation (ablation) | CMKD · KITTI val Car Moderate | 3D AP: 16.33 -> 21.54 | Improved |
| add feature distillation loss L_feat (ablation) | CMKD · KITTI val Car Easy | 3D AP: 18.81 -> 21.72 | Improved |
| add domain adaptation module (ablation) | CMKD · KITTI val Car Easy | 3D AP: 21.72 -> 23.53 | Improved |
| add IoU-aware confidence weighting to response distillation (ablation) | CMKD · KITTI val Car Moderate (KITTI train) | 3D AP: 15.69 -> 16.33 | Improved |
| use feature distillation instead of depth supervision with small training set (ablation) | CMKD · KITTI val Car Moderate (KITTI train) | 3D AP: 15.20 -> 13.04 | Hurt |
| use feature distillation instead of depth supervision with larger training set (ablation) | CMKD · KITTI val Car Moderate (KITTI train + Eigen clean) | 3D AP: 19.76 -> 21.32 | Improved |
| add depth supervision on top of feature distillation (larger set) (ablation) | CMKD · KITTI val Car Moderate (KITTI train + Eigen clean) | 3D AP: 21.32 -> 21.53 | No clear effect |
| use soft labels instead of hard labels for Pedestrian (ablation) | CMKD · KITTI val Pedestrian Moderate | 3D AP: 8.65 -> 3.20 | Hurt |
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.
- M3D-RPN — M3D-RPN: Monocular 3D Region Proposal Network for Object Detection (ICCV 2019) · 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
- SCKD: Semi-Supervised Cross-Modality Knowledge Distillation for 4D Radar Object Detection (AAAI 2025 · similarity 0.90)
- CMDA: Cross-Modal and Domain Adversarial Adaptation for LiDAR-Based 3D Object Detection (AAAI 2024 · similarity 0.89)
- SimDistill: Simulated Multi-Modal Distillation for BEV 3D Object Detection (AAAI 2024 · similarity 0.89)
- CRKD: Enhanced Camera-Radar Object Detection with Cross-modality Knowledge Distillation (CVPR 2024 · similarity 0.89)
- MonoTAKD: Teaching Assistant Knowledge Distillation for Monocular 3D Object Detection (CVPR 2025 · similarity 0.89)
- Weakly Supervised Monocular 3D Detection with a Single-View Image (CVPR 2024 · similarity 0.89)
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.
