Research map / Detection and segmentation / 3D detection and driving
Tri-Modal Fusion Transformers for UAV-based Object Detection
Craig Iaboni, Pramod Abichandani
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 MAGE gating to BiTE (ablation) | MiT-B1 fusion block · Tri-modal UAV dataset | mAP: 76.88 -> 84.24 | Improved |
| add BiTE token exchange to MAGE (ablation) | MiT-B1 fusion block · Tri-modal UAV dataset | mAP: 81.01 -> 84.24 | Improved |
| scale backbone from MiT-B1 to MiT-B4 | MAGE+BiTE · Tri-modal UAV dataset | mAP: 84.24 -> 79.97 | Hurt |
| scale backbone from MiT-B0 to MiT-B1 | MAGE+BiTE · Tri-modal UAV dataset | mAP: 80.63 -> 84.24 | Improved |
| add event modality to RGB+Thermal (ablation) | MAGE+BiTE MiT-B1 · Tri-modal UAV dataset | mAP: 83.42 -> 84.24 | Improved |
| add thermal to RGB+Event (ablation) | MAGE+BiTE MiT-B1 · Tri-modal UAV dataset | mAP: 66.32 -> 84.24 | Improved |
| use GAFF at s4 instead of s1234 (ablation) | GAFF · Tri-modal UAV dataset | mAP: 81.93 -> 83.41 | Improved |
| use multi-stage CSSA s1234 instead of s1 (ablation) | CSSA tau=0.5 · Tri-modal UAV dataset | mAP: 83.44 -> 80.91 | Hurt |
| vary CSSA threshold tau at s1 (0.3, 0.5, 0.7) (ablation) | CSSA s1 · Tri-modal UAV dataset | mAP: 83.23, 83.44, 82.98 | No clear effect |
| CSSA threshold 0.5 vs 0.3 at s3 (ablation) | CSSA s3 · Tri-modal UAV dataset | mAP: 79.83 -> 82.80 | Improved |
| tune GAFF SE ratio, guidance and merge at s4 (ablation) | GAFF · Tri-modal UAV dataset | mAP: 82.15 -> 83.78 | Mixed |
| train on day+night instead of day-only (ablation) | MiT-B1 tri-modal · Tri-modal UAV dataset | mAP: 79.0 -> 82.24 | Improved |
Closest papers by meaning
- Efficient RGB-T Object Detection via Sparse Cross-Modality Fusion (ECCV 2026 · similarity 0.88)
- Infrared-Privileged UAV Detection via Cross-Modal Vector-Quantization (AAAI 2026 · similarity 0.88)
- CCF: Complementary Collaborative Fusion for Domain Generalized Multi-Modal 3D Object Detection (CVPR 2026 · similarity 0.88)
- GAFusion: Adaptive Fusing LiDAR and Camera with Multiple Guidance for 3D Object Detection (CVPR 2024 · similarity 0.86)
- SimDistill: Simulated Multi-Modal Distillation for BEV 3D Object Detection (AAAI 2024 · similarity 0.86)
- Cocoon: Robust Multi-Modal Perception with Uncertainty-Aware Sensor Fusion (ICLR 2025 · 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.
