Object Detection (RGB-T) on M3FD
88.7Precision (People)SFEDet
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SFEDetBackbone=CSP-S, Params.=24.5M, FLOPs=112.6G2026.06 | 88.7 | 96.7 | 91.2 | 88.2 | 91 | 83.2 | 61 | 89.8 | |
| RDMIBackbone=CSP-L, Params.=72.0M, FLOPs=320G2026.06 | 84.3 | 85.4 | 87.7 | 68 | 79.6 | 70.4 | 49.7 | 79.2 | |
| Fu-MambaBackbone=CSP-L, Params.=288M, FLOPs=1133G†2026.06 | 84.3 | 94.2 | 92.9 | 80.5 | 87.5 | 88.8 | 61.9 | 88 | |
| MMFNBackbone=Res-50, Params.=176M2026.06 | 83 | 92.1 | 93.2 | 73.7 | 87.6 | 87.4 | — | 86.2 | |
| EI2DetBackbone=CSP-L, Params.=116M, FLOPs=391G†2026.06 | 82.8 | 91.6 | 91.4 | 77.2 | 84.7 | 89.3 | 55.5 | 86.2 | |
| ICAFusionBackbone=CSP-L, Params.=120M, FLOPs=370G†2026.06 | 82.3 | 93.9 | 93.4 | 89.6 | 88.2 | 87.8 | — | 88.2 | |
| EMEBackbone=Res-50, Params.=36.4M, FLOPs=316G†2026.06 | 79.5 | 89.8 | 91.9 | 74.87 | 77.4 | 84 | 54 | 82.9 |