Research map / Detection and segmentation / CNN features
"Translation, Scale and Rotation: Cross-Modal Alignment Meets RGB-Infrared Vehicle Detection"
Maoxun Yuan, Yinyan Wang, Xingxing Wei
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 alignment process (position, size, angle) to two-stream baseline (ablation) | Two-stream Oriented R-CNN · DroneVehicle val | mAP: 68.62 -> 71.13 | Improved |
| add size branch to position-only alignment (ablation) | TSRA alignment process · DroneVehicle val | mAP: 69.49 -> 70.71 | Improved |
| add angle branch to position-only alignment (ablation) | TSRA alignment process · DroneVehicle val | mAP: 69.49 -> 70.27 | Improved |
| add angle branch to position+size alignment (ablation) | TSRA alignment process · DroneVehicle val | mAP: 70.71 -> 71.13 | Improved |
| add Modality-Selection strategy to alignment (ablation) | TSFADet · DroneVehicle val | mAP: 71.13 -> 72.36 | Improved |
| add Multi-task Jitter to alignment with MS (ablation) | TSFADet · DroneVehicle val | mAP: 72.05 -> 73.06 | Improved |
| add Multi-task Jitter to alignment with MS (w/o MJ vs full) (ablation) | TSFADet · DroneVehicle val | mAP: 72.36 -> 73.06 | Improved |
| use MS strategy instead of RGB modality as reference (ablation) | TSFADet · DroneVehicle val | mAP: 71.57 -> 73.06 | Improved |
| use MS strategy instead of IR modality as reference (ablation) | TSFADet · DroneVehicle val | mAP: 72.05 -> 73.06 | Improved |
| use MS strategy instead of random reference selection (ablation) | TSFADet · DroneVehicle val | mAP: 72.42 -> 73.06 | Improved |
| add Cascade R-CNN structure to TSFADet | TSFADet · DroneVehicle val | mAP: 73.06 -> 73.90 | Improved |
| use RGB+IR input with TSFADet instead of IR-only Oriented R-CNN | Oriented R-CNN · DroneVehicle val | mAP: 65.50 -> 73.06 | Improved |
Closest papers by meaning
- Infrared-Privileged UAV Detection via Cross-Modal Vector-Quantization (AAAI 2026 · similarity 0.89)
- Weakly Misalignment-free Adaptive Feature Alignment for UAVs-based Multimodal Object Detection (CVPR 2024 · similarity 0.89)
- IGIANet: Illumination Guided Implicit Alignment Network for Infrared–Visible UAV Detection (AAAI 2026 · similarity 0.87)
- Weakly Aligned Cross-Modal Learning for Multispectral Pedestrian Detection (ICCV 2019 · similarity 0.87)
- Alignment-Free RGB-T Salient Object Detection: A Large-Scale Dataset and Progressive Correlation Network (AAAI 2025 · similarity 0.86)
- FRED: Towards a Full Rotation-Equivariance in Aerial Image Object Detection (AAAI 2024 · similarity 0.85)
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.
