Object Detection on LLVIP
98.3mAP50waveMamba
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| waveMambaModality=RGB+IR, Year=2025, Backbone=CSPDarknet, Params/M=69.12026.04 | 98.3 | 66 | — | — | 57.1 | — | — | |
| SLGNetModality=RGB+IR, Year=2026, Backbone=ViT-based, Params/M=12.12026.04 | 98.3 | 66.1 | 75.4 | — | 55.6 | — | — | |
| Semantic Bridge FusionModality=RGB+IR, Year=2026, Backbone=CSPDarknet, Params/M=15.52026.04 | 98.3 | 68.9 | 81 | — | 59.7 | — | — | |
| MiPa2024.04 | 98.2 | 66.5 | 78.1 | — | — | — | — | |
| EI2Det2026.03 | 98 | — | — | — | — | — | — | |
| MDACLData Type=IR+RGB, Backbone=CSPDarknet532026.01 | 97.9 | 66.5 | — | — | — | — | — | |
| MDACLModality=RGB+IR, Year=2026, Backbone=CSPDarknet2026.04 | 97.9 | 66.5 | — | — | 55.6 | — | — | |
| G-DINO (FT)Evaluation Protocol=fine-tuned2026.05 | 97.9 | 70.6 | — | — | — | — | — | |
| DHANet2026.03 | 97.7 | — | — | — | — | — | — | |
| CrossModalNetModality=RGB+IR, Year=2025, Backbone=CSPDarknet, Params/M=92.82026.04 | 97.7 | 64.7 | 73.5 | — | 54 | — | — | |
| COFNetModality=RGB+IR, Year=2025, Backbone=CSPDarknet, Params/M=90.22026.04 | 97.7 | 65.9 | 75.9 | — | 55.3 | — | — | |
| FFMData Type=IR+RGB, Backbone=CSPDarknet532026.01 | 97.6 | 64.8 | — | — | — | — | — | |
| MCORModality=RGB+IR, Year=2025, Backbone=CSPDarknet2026.04 | 97.6 | 64.9 | 73.6 | — | 52.4 | — | — | |
| CFT2024.04 | 97.5 | 63.6 | 72.9 | — | — | — | — | |
| DLRMamba2026.03 | 97.5 | — | — | — | — | — | — | |
| MS2FusionModality=RGB+IR, Year=2026, Backbone=CSPDarknet, Params/M=130.22026.04 | 97.5 | 65.5 | — | — | 52.9 | — | — | |
| CrossFormer2024.04 | 97.4 | 65.1 | 75.4 | — | — | — | — | |
| Fusion-MambaData Type=IR+RGB, Backbone=CSPDarknet532026.01 | 97 | 64.3 | — | — | — | — | — | |
| Fusion MambaModality=RGB+IR, Year=2025, Backbone=CSPDarknet, Params/M=287.62026.04 | 97 | 64.3 | — | — | 55.7 | — | — | |
| PSFusion2026.03 | 96.6 | — | — | — | — | — | — | |
| YOLO-Adaptor2026.03 | 96.5 | — | — | — | — | — | — | |
| UniRGB-IRModality=RGB+IR2024.04 | 96.1 | 63.2 | 72.2 | — | — | — | — | |
| UniRGB-IRData Type=IR+RGB, Backbone=Transformer2026.01 | 96.1 | 63.2 | — | — | — | — | — | |
| PIAFusion2026.03 | 96.1 | — | — | — | — | — | — | |
| InfraredBackbone=YOLOv8, Training Epochs=3002024.12 | 96.03 | — | 72.07 | — | — | — | — | |
| M2I2HAModality=RGB+IR, Backbone=YOLOv8s, GPU-H (h)=2.72026.01 | 95.9 | 58.8 | 65.6 | — | — | — | — | |
| RSDetModality=RGB+IR2024.04 | 95.8 | 61.3 | 70.4 | — | — | — | — | |
| RSDet2024.04 | 95.8 | 61.3 | — | — | — | — | — | |
| RSDetModality=RGB+IR, Year=2024, Backbone=ResNet502026.04 | 95.8 | 61.3 | 70.4 | — | 52.6 | — | — | |
| CDDFuse2026.03 | 95.7 | — | — | — | — | — | — | |
| COMOModality=RGB+IR, Backbone=YOLOv8s, GPU-H (h)=22026.01 | 95.5 | 57.3 | 62 | — | — | — | — | |
| M2I2HAModality=RGB+IR, Backbone=YOLOv8n, GPU-H (h)=1.92026.01 | 95.5 | 59.7 | 68 | — | — | — | — | |
| IFCNN2026.03 | 95.5 | — | — | — | — | — | — | |
| YOLO Fusion2026.03 | 95.4 | — | — | — | — | — | — | |
| M2I2HAModality=RGB+IR, Backbone=YOLOv13s, GPU-H (h)=4.32026.01 | 95.3 | 58.9 | 65.5 | — | — | — | — | |
| CFTModality=RGB+IR, Backbone=YOLOv5s, GPU-H (h)=1.42026.01 | 95.2 | 60 | 66.5 | — | — | — | — | |
| ICAFusionData Type=IR+RGB, Backbone=CSPDarknet532026.01 | 95.2 | 60.1 | — | — | — | — | — | |
| YOLOv5Modality=IR, Year=2020, Backbone=CSPDarknet, Params/M=47.62026.04 | 95.2 | 62.1 | 67 | — | 50.2 | — | — | |
| TDFusionBackbone=YOLOv8, Training Epochs=3002024.12 | 95 | — | 69.18 | — | — | — | — | |
| DCINNBackbone=YOLOv8, Training Epochs=3002024.12 | 94.92 | — | 68.34 | — | — | — | — | |
| Cascade R-CNN (ViT-B)Modality=IR2024.04 | 94.9 | 60.4 | 67.5 | — | — | — | — | |
| YOLOv8-IRModality=IR, Backbone=YOLOv8s, GPU-H (h)=-2026.01 | 94.9 | 59.6 | 66.1 | — | — | — | — | |
| RetinaNetModality=IR2024.04 | 94.8 | 55.1 | 57.6 | — | — | — | — | |
| RetinaNetModality=IR, Year=2017, Backbone=ResNet50, Params/M=43.02026.04 | 94.8 | 55.1 | 57.6 | — | 43.3 | — | — | |
| GM-DETRModality=RGB+IR, Backbone=RT-DETR, GPU-H (h)=7.12026.01 | 94.7 | 55.7 | 60.2 | — | — | — | — | |
| M2I2HAModality=RGB+IR, Backbone=YOLOv13n, GPU-H (h)=2.52026.01 | 94.7 | 55.4 | 59.2 | — | — | — | — | |
| Faster R-CNNModality=IR2024.04 | 94.6 | 54.5 | 57.6 | — | — | — | — | |
| YOLOv8l-IRData Type=IR, Backbone=CSPDarknet532026.01 | 94.6 | 61.7 | — | — | — | — | — | |
| FasterRCNNModality=IR, Year=2015, Backbone=ResNet50, Params/M=41.12026.04 | 94.6 | 54.5 | 57.6 | — | 46.1 | — | — | |
| ICAFusionModality=RGB+IR, Backbone=YOLOv5s, GPU-H (h)=7.22026.01 | 94.5 | 57.7 | 63.1 | — | — | — | — | |
| CSAAModality=RGB+IR2024.04 | 94.3 | 59.2 | 66.6 | — | — | — | — | |
| CSSA2024.04 | 94.3 | 59.2 | 66.6 | — | — | — | — | |
| CSAAData Type=IR+RGB, Backbone=ResNet502026.01 | 94.3 | 54.2 | — | — | — | — | — | |
| MURFBackbone=YOLOv8, Training Epochs=3002024.12 | 94.24 | — | 68.04 | — | — | — | — | |
| YOLOv13-IRModality=IR, Backbone=YOLOv13s, GPU-H (h)=-2026.01 | 94.2 | 57 | 63.2 | — | — | — | — | |
| LUT-FuseData Type=IR+RGB, Backbone=CSPDarknet532026.01 | 94.1 | 61.4 | — | — | — | — | — | |
| Text-IFData Type=IR+RGB, Backbone=Transformer2026.01 | 94.1 | 60.2 | — | — | — | — | — | |
| GAFFModality=RGB+IR2024.04 | 94 | 55.8 | 60.2 | — | — | — | — | |
| GAFF2024.04 | 94 | 55.8 | 60.2 | — | — | — | — | |
| EMMABackbone=YOLOv8, Training Epochs=3002024.12 | 94 | — | 66.21 | — | — | — | — | |
| SegMIFBackbone=YOLOv8, Training Epochs=3002024.12 | 93.95 | — | 66.45 | — | — | — | — | |
| DDQ-DETRModality=IR2024.04 | 93.9 | 58.6 | 64.6 | — | — | — | — | |
| DDQ-DETRModality=IR, Year=2023, Backbone=ResNet50, Params/M=244.62026.04 | 93.9 | 58.6 | 64.6 | — | 47.9 | — | — | |
| TarDALBackbone=YOLOv8, Training Epochs=3002024.12 | 93.79 | — | 62.71 | — | — | — | — | |
| TIMFusionBackbone=YOLOv8, Training Epochs=3002024.12 | 93.76 | — | 61.33 | — | — | — | — | |
| GHOSTModality=RGB+IR, Backbone=YOLOv5s, GPU-H (h)=3.62026.01 | 93.7 | 58.9 | 62.5 | — | — | — | — | |
| ProbEnModality=RGB+IR2024.04 | 93.4 | 51.5 | 50.2 | — | — | — | — | |
| ProbEn2024.04 | 93.4 | 51.5 | 50.2 | — | — | — | — | |
| MRFSBackbone=YOLOv8, Training Epochs=3002024.12 | 93.03 | — | 67.21 | — | — | — | — | |
| YOLO v8l-Worldv2Similarity-driven re-scoring (BEM)=true2026.04 | 92.36 | — | — | 86.27 | — | — | — | |
| YOLOv5Modality=RGB, Year=2020, Backbone=CSPDarknet, Params/M=47.62026.04 | 91.9 | 54 | 52.5 | — | 41.1 | — | — | |
| YOLOv8l-RGBData Type=RGB, Backbone=CSPDarknet532026.01 | 91.8 | 53.6 | — | — | — | — | — | |
| VisibleBackbone=YOLOv8, Training Epochs=3002024.12 | 91.78 | — | 48.66 | — | — | — | — | |
| Halfway Fusion2024.04 | 91.4 | 55.1 | 60.1 | — | — | — | — | |
| YOLO v8s-Worldv2Similarity-driven re-scoring (BEM)=true2026.04 | 91.36 | — | — | 81.88 | — | — | — | |
| YOLO v8l-Worldv2Similarity-driven re-scoring (BEM)=false2026.04 | 91.2 | — | — | 86.22 | — | — | — | |
| SuperYOLOModality=RGB+IR, Backbone=YOLOv5s, GPU-H (h)=2.82026.01 | 91 | 54 | 58.1 | — | — | — | — | |
| MoE-Fusion2026.03 | 91 | — | — | — | — | — | — | |
| Cascade R-CNN (ViT-B)Modality=RGB2024.04 | 90.3 | 51.7 | 54.7 | — | — | — | — | |
| YOLO v8s-Worldv2Similarity-driven re-scoring (BEM)=false2026.04 | 90.23 | — | — | 81.78 | — | — | — | |
| SSDModality=IR2024.04 | 90.2 | 53.5 | 57.9 | — | — | — | — | |
| DIVFusion2024.04 | 89.8 | 52 | — | — | — | — | — | |
| DIVFusion2026.03 | 89.8 | — | — | — | — | — | — | |
| YOLOv8-RGBModality=RGB, Backbone=YOLOv8s, GPU-H (h)=-2026.01 | 88.7 | 49.8 | 51.5 | — | — | — | — | |
| DM-Fusion2026.03 | 88.1 | — | — | — | — | — | — | |
| RetinaNetModality=RGB2024.04 | 88 | 42.8 | 34.4 | — | — | — | — | |
| RetinaNetModality=RGB, Year=2017, Backbone=ResNet50, Params/M=43.02026.04 | 88 | 42.8 | 34.4 | — | 32.4 | — | — | |
| Faster R-CNNModality=RGB2024.04 | 87 | 45.1 | 41.2 | — | — | — | — | |
| FasterRCNNModality=RGB, Year=2015, Backbone=ResNet50, Params/M=41.12026.04 | 87 | 45.1 | 41.2 | — | 36.4 | — | — | |
| DDQ-DETRModality=RGB2024.04 | 86.1 | 46.7 | 45.8 | — | — | — | — | |
| YOLOv13-RGBModality=RGB, Backbone=YOLOv13s, GPU-H (h)=-2026.01 | 86.1 | 46.3 | 44 | — | — | — | — | |
| DDQ-DETRModality=RGB, Year=2023, Backbone=ResNet50, Params/M=244.62026.04 | 86.1 | 46.7 | 45.8 | — | 38.8 | — | — | |
| Thermal-DetEvaluation Protocol=zero-shot2026.05 | 85.6 | 56.6 | — | — | — | — | — | |
| SSDModality=RGB2024.04 | 82.6 | 39.8 | 31.8 | — | — | — | — | |
| YOLO v11m (COCO)Similarity-driven re-scoring (BEM)=true2026.04 | 80.99 | — | — | 92.87 | — | — | — | |
| YOLO v11m (COCO)Similarity-driven re-scoring (BEM)=false2026.04 | 80.49 | — | — | 89.82 | — | — | — | |
| MM-GDINOEvaluation Protocol=zero-shot2026.05 | 80 | 56.1 | — | — | — | — | — | |
| RT-DETR-L (COCO)Similarity-driven re-scoring (BEM)=true2026.04 | 79.59 | — | — | 82.85 | — | — | — | |
| RT-DETR-L (COCO)Similarity-driven re-scoring (BEM)=false2026.04 | 79.26 | — | — | 77.6 | — | — | — | |
| YOLO v8s (COCO)Similarity-driven re-scoring (BEM)=true2026.04 | 75.9 | — | — | 91.63 | — | — | — |