Research map / Detection and segmentation / 3D detection and driving
CramNet: Camera-Radar Fusion with Ray-Constrained Cross-Attention for Robust 3D Object Detection
Jyh-Jing Hwang, Henrik Kretzschmar, Joshua Manela, Sean Rafferty, Nicholas Armstrong-Crews, Tiffany Chen, Dragomir Anguelov
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 ray-constrained cross-attention (ablation) | camera-radar fusion · RADIATE (overall) | BEV AP: 56.19 -> 60.20 | Improved |
| add sensor dropout (ablation) | camera-radar fusion · RADIATE (overall) | BEV AP: 56.19 -> 61.23 | Improved |
| add sensor dropout to cross-attention model (ablation) | CramNet · RADIATE (overall) | BEV AP: 60.20 -> 62.07 | Improved |
| add cross-attention to sensor-dropout model (ablation) | CramNet · RADIATE (overall) | BEV AP: 61.23 -> 62.07 | Improved |
| add modality code to point features (ablation) | CramNet · RADIATE (overall) | BEV AP: 53.4 -> 62.1 | Improved |
| fuse camera with radar instead of radar only | CramNet · RADIATE (overall) | BEV AP: 56.19 -> 62.07 | Improved |
| use CramNet fusion instead of Faster R-CNN baseline | CramNet · RADIATE (overall) | BEV AP: 46.55 -> 62.07 | Improved |
| sparsify radar RF image (intensity threshold 0.25) (ablation) | CramNet · RADIATE (overall) | BEV AP: 60.20 -> 50.90 | Hurt |
| sparsify radar RF image (intensity threshold 0.5) (ablation) | CramNet · RADIATE (overall) | BEV AP: 60.20 -> 17.81 | Hurt |
| train with sensor dropout, evaluate on camera images with Gaussian noise (ablation) | CramNet · RADIATE (camera noise) | BEV AP: 2 to 10 percentage points | Improved |
| apply conventional dropout on point features instead of sensor-wise dropout (ablation) | CramNet · RADIATE (overall) | BEV AP | No clear effect |
| increase sensor dropout probability beyond 0.2 (ablation) | CramNet · RADIATE (overall) | BEV AP | 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 2 of 2 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
- CRT-Fusion: Camera, Radar, Temporal Fusion Using Motion Information for 3D Object Detection (NeurIPS 2024 · similarity 0.90)
- RobuRCDet: Enhancing Robustness of Radar-Camera Fusion in Bird's Eye View for 3D Object Detection (ICLR 2025 · similarity 0.88)
- RCBEVDet: Radar-camera Fusion in Bird's Eye View for 3D Object Detection (CVPR 2024 · similarity 0.88)
- CRKD: Enhanced Camera-Radar Object Detection with Cross-modality Knowledge Distillation (CVPR 2024 · similarity 0.88)
- RICCARDO: Radar Hit Prediction and Convolution for Camera-Radar 3D Object Detection (CVPR 2025 · similarity 0.88)
- RCTrans: Radar-Camera Transformer via Radar Densifier and Sequential Decoder for 3D Object Detection (AAAI 2025 · similarity 0.87)
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.
