Object Detection on FLIR Aligned (test)
87mAP@0.5WD-FQDet
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| WD-FQDetBackbone=ResNet50, Modal=IR+RGB2026.05 | 87 | 89.3 | 79.3 | 92.4 | 50.1 | 50.2 | |
| Scene-adaptive CBAMInference Speed (s)=0.0322023.10 | 86.16 | 88.92 | 78.61 | 90.94 | 43 | 47.1 | |
| Scene-agnostic CBAMInference Speed (s)=0.0282023.10 | 85.45 | 88.26 | 77.43 | 90.68 | 43.3 | 46.8 | |
| FD2NetVenue=AAAI’25, Backbone=ResNet50, Modal=IR+RGB2026.05 | 82.9 | 85.3 | 73.2 | 89.9 | 42.5 | — | |
| GM-DETR†Venue=CVPR’24, Backbone=ResNet50, Modal=IR+RGB2026.05 | 81.9 | 87.9 | 66.6 | 91.3 | 42.8 | 45.4 | |
| TFDetVenue=TNNLS’24, Backbone=CSP53, Modal=IR+RGB2026.05 | 81.7 | 85.2 | 71.9 | 87.5 | 41.3 | 46.6 | |
| SegMiFVenue=ICCV’23, Backbone=ViT, Modal=IR+RGB2026.05 | 81.5 | 85.3 | 72.8 | 86.9 | 40.9 | — | |
| MetaFVenue=CVPR’23, Modal=IR+RGB2026.05 | 81.4 | 83.3 | 71.1 | 89.2 | 40.7 | — | |
| DDFMVenue=ICCV’23, Modal=IR+RGB2026.05 | 81.2 | 84.5 | 71.5 | 87.9 | 40.2 | — | |
| CDDFuseVenue=CVPR’23, Backbone=Restormer, Modal=IR+RGB2026.05 | 80.8 | 82.3 | 72.9 | 87.2 | 39.4 | — | |
| LRAF-NetVenue=TNNLS’23, Backbone=CSP53, Modal=IR+RGB2026.05 | 80.6 | 83.3 | 69.7 | 88.8 | 41 | 42.8 | |
| LRAF-Net2023.10 | 80.5 | — | — | — | — | 42.8 | |
| RT-DETR†Venue=CVPR’24, Backbone=ResNet50, Modal=IR2026.05 | 80.2 | 83.8 | 66.6 | 90.2 | 39.2 | 43.2 | |
| FasterRCNN + MFPTInference Speed (s)=0.082023.10 | 80 | 83.2 | 67.7 | 89 | — | — | |
| TarDALVenue=CVPR’22, Backbone=CSPD53, Modal=IR+RGB2026.05 | 79.9 | 85.1 | 69.3 | 85.3 | 37.9 | — | |
| CSSAVenue=CVPR’23, Backbone=ResNet50, Modal=IR+RGB2026.05 | 79.4 | 83.2 | 68.6 | 86.7 | 37.2 | 41.3 | |
| CSAAInference Speed (s)=0.0312023.10 | 79.2 | — | — | — | 37.4 | 41.3 | |
| CFTInference Speed (s)=0.0262023.10 | 78.7 | — | — | — | 35.5 | 40.2 | |
| UA-CMDet2023.10 | 78.6 | 83.2 | 64.3 | 88.4 | — | — | |
| SDNetVenue=IJCV’21, Modal=IR+RGB2026.05 | 77.5 | 81 | 64.2 | 87.3 | 33.1 | — | |
| RetinaNet + MFPTInference Speed (s)=0.052023.10 | 76.8 | 78.1 | 65 | 87.3 | — | — | |
| CFR-3Inference Speed (s)=0.052023.10 | 72.93 | 74.49 | 57.77 | 84.91 | — | — | |
| GAFFInference Speed (s)=0.0612023.10 | 72.9 | 76.6 | 59.4 | 85.5 | 32.9 | 37.5 | |
| M2I2HA (YOLOv8s)Modality=RGB+IR, Backbone=YOLOv8s2026.01 | 72.5 | 81 | 47.7 | 88.9 | 33.4 | 37.8 | |
| Thermal onlyInference Speed (s)=0.0162023.10 | 72.16 | 82.86 | 50.8 | 82.83 | 33.4 | 37 | |
| COMOModality=RGB+IR, Backbone=YOLOv8s2026.01 | 71.7 | 77.8 | 49.4 | 87.9 | 34 | 37.5 | |
| M2I2HA (YOLOv8n)Modality=RGB+IR, Backbone=YOLOv8n2026.01 | 71.3 | 78.9 | 45.6 | 89.3 | 33.1 | 37.1 | |
| YOLOv8-IRModality=IR, Backbone=YOLOv8s2026.01 | 70.8 | 77.1 | 47.7 | 87.7 | 32.4 | 36.5 | |
| SuperYOLOModality=RGB+IR, Backbone=YOLOv5s2026.01 | 70.8 | 77 | 46.9 | 88.5 | 29.3 | 34.1 | |
| GM-DETRModality=RGB+IR, Backbone=RT-DETR2026.01 | 70.2 | 76.2 | 46.1 | 88.3 | 29.1 | 34.6 | |
| GHOSTModality=RGB+IR, Backbone=YOLOv5s2026.01 | 68.9 | 75.6 | 46.6 | 84.5 | 28.4 | 30.6 | |
| RT-DETR†Venue=CVPR’24, Backbone=ResNet50, Modal=RGB2026.05 | 68.3 | 69.2 | 53.6 | 82.1 | 26 | 33 | |
| CFTModality=RGB+IR, Backbone=YOLOv5s2026.01 | 68.2 | 73.5 | 44.9 | 86.2 | 27.5 | 32.6 | |
| M2I2HA (YOLOv13s)Modality=RGB+IR, Backbone=YOLOv13s2026.01 | 67.4 | 74.2 | 41.3 | 86.6 | 30.2 | 34.8 | |
| M2I2HA (YOLOv13n)Modality=RGB+IR, Backbone=YOLOv13n2026.01 | 66.8 | 74.3 | 39.1 | 86.9 | 28.9 | 34 | |
| ICAFusionModality=RGB+IR, Backbone=YOLOv5s2026.01 | 66.4 | 74 | 37.2 | 88.1 | 25.2 | 31.3 | |
| YOLOv13-IRModality=IR, Backbone=YOLOv13s2026.01 | 61.7 | 68.5 | 31.8 | 84.7 | 26.6 | 31.3 | |
| YOLOv8-RGBModality=RGB, Backbone=YOLOv8s2026.01 | 58 | 57 | 36.9 | 79.5 | 20.2 | 26.4 | |
| RGB onlyInference Speed (s)=0.0162023.10 | 57.32 | 60.79 | 37.25 | 73.94 | 17.6 | 24.7 | |
| YOLOv13-RGBModality=RGB, Backbone=YOLOv13s2026.01 | 48.7 | 48.6 | 22.6 | 74.6 | 17.4 | 22.6 |