Semantic Segmentation on URTF (val)
71.04mIoUGSCNet
Evaluation Results
| Method | Links | |
|---|---|---|
| GSCNetBackbone=MiT-B4, Params/M=160.66, FLOPs/G=208.07, FPS=16.62, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 71.04 | |
| MambaSegBackbone=VMamba-T, Params/M=315.44, FLOPs/G=186.06, FPS=20.92, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 68.31 | |
| AMDANetBackbone=MiT-B4, Params/M=135.76, FLOPs/G=213.43, FPS=11.03, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 68.13 | |
| CRMBackbone=MiT-B4, Params/M=193.51, FLOPs/G=267.48, FPS=15.60, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 67.78 | |
| MRFSBackbone=MiT-B4, Params/M=134.99, FLOPs/G=140.65, FPS=19.55, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 65.28 | |
| DFormerV2Backbone=DFormerV2-L, Params/M=95.57, FLOPs/G=146.88, FPS=17.53, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 64.94 | |
| Mul-VMambaBackbone=VMamba-T, Params/M=112.23, FLOPs/G=56.33, FPS=40.40, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 64.92 | |
| CMNextBackbone=MiT-B4, Params/M=116.59, FLOPs/G=154.89, FPS=24.90, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 62.78 | |
| SGFNetBackbone=Res50, Params/M=125.31, FLOPs/G=157.25, FPS=24.04, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 62.41 | |
| GeminiFusionBackbone=MiT-B4, Params/M=103.32, FLOPs/G=271.75, FPS=14.20, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 61.73 | |
| MiLNetBackbone=MiT-B4, Params/M=126.04, FLOPs/G=197.33, FPS=22.40, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 61.29 | |
| CMXBackbone=MiT-B4, Params/M=143.38, FLOPs/G=172.27, FPS=22.80, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 60.31 | |
| SegFormerBackbone=MiT-B4, Params/M=61.39, FLOPs/G=76.18, FPS=44.68, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-only2026.04 | 60.05 | |
| EFENetBackbone=Twins-L, Params/M=143.84, FLOPs/G=198.99, FPS=22.01, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-only2026.04 | 59.83 | |
| Mask2FormerBackbone=Swin-B, Params/M=107.00, FLOPs/G=172.00, FPS=24.80, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-only2026.04 | 59.64 | |
| SegNeXtBackbone=MSCAN-L, Params/M=48.84, FLOPs/G=82.19, FPS=53.15, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-only2026.04 | 59.22 | |
| HRNetBackbone=HRNet-W48, Params/M=65.89, FLOPs/G=117.80, FPS=51.75, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-only2026.04 | 56.63 | |
| ASANetBackbone=ConvNeXt-T, Params/M=82.93, FLOPs/G=129.80, FPS=43.20, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 53.91 | |
| DeepLabv3+Backbone=Res101, Params/M=60.24, FLOPs/G=318.00, FPS=44.49, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-only2026.04 | 53.63 | |
| EGFNetBackbone=Res152, Params/M=62.93, FLOPs/G=230.14, FPS=19.62, Resolution=640 × 512, Batch size=1, Hardware=NVIDIA H20 GPU, Modality=RGB-T Fusion2026.04 | 48.74 |