Infrared Small Target Detection on IRSTD-1K
96.9PdInfMAE
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| InfMAEBackbone=ViT-B, Model=IRSTD2025.12 | 96.9 | — | — | — | 66.5 | — | — | — | — | |
| SAISTPublish=CVPR’252026.05 | 96.18 | 72.14 | 4.76 | — | — | — | — | — | — | |
| MDAFNetPublication=-2026.01 | 95.92 | 70.11 | 8.43 | — | — | — | — | — | — | |
| DuGI-MAEBackbone=ViT-B, Model=IRSTD2025.12 | 95.9 | — | — | — | 67.1 | — | — | — | — | |
| IRMambaPublish=AAAI’252026.05 | 95.81 | 70.04 | 5.92 | — | — | — | — | — | — | |
| DCCS-DetType=CNN-Mamba, Publication=-2026.01 | 95.58 | 69.64 | — | 10.48 | — | — | — | — | — | |
| ISNetPublish=CVPR’222026.05 | 95.56 | 68.77 | 15.39 | — | — | — | — | — | — | |
| GSACP-FinalZero-shot transfer=true, Train dataset=SIRST3, Source threshold=0.552026.05 | 95.51 | — | 16.1 | — | 63.58 | 63.26 | — | -0.0924 | 65.19 | |
| Na-IRSTD2026.05 | 95.42 | 73.42 | 4.33 | — | — | — | — | — | — | |
| DNANetPublish=TIP’222026.05 | 94.95 | 68.87 | 13.38 | — | — | — | — | — | — | |
| ECFNetFLOPs=34.3, Params=24.1, FPS=852026.06 | 94.91 | — | 9.94 | — | — | 69.67 | — | — | — | |
| MSHNetScale loss variant=L3 (Mobius), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 94.9 | — | 12.68 | — | 68.28 | — | — | — | — | |
| ISNetDescription=CNN Full2023.04 | 94.28 | 60.61 | 61.28 | — | — | — | — | — | — | |
| UIUNetPublish=TIP’232026.05 | 94.27 | 69.13 | 16.47 | — | — | — | — | — | — | |
| AGPCNetPublish=TAES’232026.05 | 94.26 | 68.81 | 15.85 | — | — | — | — | — | — | |
| MSHNetScale loss variant=L1 (Diff-based), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 94.22 | — | 7.67 | — | 69.19 | — | — | — | — | |
| MSHNetScale loss variant=L1 (Diff-based)2026.04 | 94.22 | — | 16.02 | — | 67.71 | — | — | — | — | |
| UIU-Net2026.04 | 93.98 | — | 22.07 | — | 66.66 | — | — | — | — | |
| MSHNetPublication=CVPR’242026.01 | 93.88 | 67.16 | 15.03 | — | — | — | — | — | — | |
| PConvPublication=AAAI’252026.01 | 93.88 | 67.58 | 12.22 | — | — | — | — | — | — | |
| MSHNetType=CNN, Publication=CVPR, 20242026.01 | 93.88 | 67.16 | — | 15.03 | — | — | — | — | — | |
| PConvType=CNN, Publication=AAAI, 20252026.01 | 93.88 | 67.58 | — | 12.22 | — | — | — | — | — | |
| REEMDescription=SCR-Aware Reweighting2026.06 | 93.88 | 68.44 | 6.3 | — | — | — | — | — | — | |
| MSHNetPub’Year=CVPR’24, FLOPs=38.2, Params=15.5, FPS=782026.06 | 93.88 | — | 15.03 | — | — | 67.16 | — | — | — | |
| MSHNetScale loss variant=L2 (Var-based)2026.04 | 93.85 | — | 15.03 | — | 67.34 | — | — | — | — | |
| ISNet2026.05 | 93.6 | 65.59 | — | 45.97 | — | 61.74 | 79.19 | — | — | |
| UIUNetStrategy=Full2026.05 | 93.6 | 70.36 | 18.24 | — | — | 64.09 | — | — | — | |
| MSHNetScale loss variant=L2 (Var-based), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 93.54 | — | 11.77 | — | 68.56 | — | — | — | — | |
| MSHNetScale loss variant=L2 (Var-based), Gaussian-shaped spatial attention (GP)=true2026.04 | 93.54 | — | 26.57 | — | 65.44 | — | — | — | — | |
| MSHNetScale loss variant=L4 (Var-denominator), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 93.54 | — | 8.45 | — | 68.25 | — | — | — | — | |
| MSHNetScale loss variant=L4 (Var-denominator), Gaussian-shaped spatial attention (GP)=true2026.04 | 93.54 | — | 8.2 | — | 67.33 | — | — | — | — | |
| DNA-NetDescription=CNN Full2023.04 | 93.27 | 62.73 | 21.81 | — | — | — | — | — | — | |
| ACM2026.04 | 93.27 | — | 65.28 | — | 57.03 | — | — | — | — | |
| MTU-Net2026.04 | 93.27 | — | 36.8 | — | 63.24 | — | — | — | — | |
| SCTransNet2026.04 | 93.27 | — | 10.74 | — | 68.15 | — | — | — | — | |
| DNANetPub’Year=TIP’22, FLOPs=89.3, Params=4.7, FPS=472026.06 | 93.27 | — | 17.61 | — | — | 62.73 | — | — | — | |
| MSHNetScale loss variant=L1 (Diff-based), Gaussian-shaped spatial attention (GP)=true2026.04 | 93.2 | — | 9.95 | — | 68.98 | — | — | — | — | |
| MSHNetScale loss variant=L2 (Var-based), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 93.2 | — | 23.76 | — | 65.3 | — | — | — | — | |
| MSHNetDescription=Deep Learning2026.06 | 93.2 | 65.6 | 13.51 | — | — | — | — | — | — | |
| UIU-NetDescription=CNN Full2023.04 | 92.93 | 61.11 | 26.87 | — | — | — | — | — | — | |
| FSGNetMethod Category=Deep Learning Methods2026.03 | 92.93 | 72.45 | 5.43 | — | — | 68.04 | — | — | — | |
| MLCLNetStrategy=Full2026.05 | 92.93 | 64.29 | 7.91 | — | — | 64.22 | — | — | — | |
| UIU-NetPub’Year=TIP’23, FLOPs=140.6, Params=54.5, FPS=512026.06 | 92.93 | — | 26.87 | — | — | 61.11 | — | — | — | |
| GSFANetPub’Year=TGRS’25, FLOPs=63.6, Params=8.8, FPS=622026.06 | 92.92 | — | 20.17 | — | — | 60.52 | — | — | — | |
| RDIANPublication=TGRS’232026.01 | 92.86 | 63.4 | 11.54 | — | — | — | — | — | — | |
| RDIANType=CNN, Publication=TGRS, 20232026.01 | 92.86 | 63.4 | — | 11.54 | — | — | — | — | — | |
| MSHNetScale loss variant=L3 (Mobius), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 92.86 | — | 11.24 | — | 67.45 | — | — | — | — | |
| MSHNetScale loss variant=L3 (Mobius)2026.04 | 92.86 | — | 18.22 | — | 66.65 | — | — | — | — | |
| AGPCNetMethod Category=Deep Learning Methods2026.03 | 92.83 | 66.29 | 13.12 | — | — | 65.23 | — | — | — | |
| HaarTransNet2026.05 | 92.62 | 66.65 | — | 34.86 | — | 64.04 | 80.03 | — | — | |
| ACMDescription=CNN Coarse+, Supervision=Coarse Point, Framework=LESPS2023.04 | 92.59 | 40.37 | 64.81 | — | — | — | — | — | — | |
| RDIAN2026.05 | 92.59 | 63.3 | — | 38.43 | — | 63.35 | 77.52 | — | — | |
| AGPCNet2026.05 | 92.59 | 65.55 | — | 44.6 | — | 62.42 | 79.18 | — | — | |
| GCLNet2026.05 | 92.59 | 65.35 | — | 28.7 | — | 63.41 | 79.04 | — | — | |
| LoHGNet2026.05 | 92.59 | 68.03 | — | 25.75 | — | 64.26 | 80.28 | — | — | |
| DNANetStrategy=Ours2026.05 | 92.59 | 63.64 | 18.67 | — | — | 64.67 | — | — | — | |
| ISNetDescription=Deep Learning2026.06 | 92.59 | 62.88 | 27.92 | — | — | — | — | — | — | |
| SCTransNetPublication=TGRS’242026.01 | 92.52 | 66.07 | 15.79 | — | — | — | — | — | — | |
| SCTransNetType=CNN-Transformer, Publication=TGRS, 20242026.01 | 92.52 | 66.07 | — | 15.79 | — | — | — | — | — | |
| MSHNetMethod Category=Deep Learning Methods2026.03 | 92.52 | 67.17 | 12.6 | — | — | 60.38 | — | — | — | |
| MSHNetScale loss variant=L1 (Diff-based), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 92.52 | — | 20.42 | — | 64.87 | — | — | — | — | |
| ALCLNetStrategy=Ours2026.05 | 92.52 | 64.72 | 18.07 | — | — | 60.35 | — | — | — | |
| RPCASSMParams (M)=0.452026.06 | 92.44 | — | — | 22.31 | 68.44 | — | 81.26 | — | — | |
| ISNetPub’Year=CVPR’22, FLOPs=185.4, Params=9.7, FPS=632026.06 | 92.33 | — | 18.57 | — | — | 64.23 | — | — | — | |
| ALCNetDescription=CNN Coarse+, Supervision=Coarse Point, Framework=LESPS2023.04 | 92.26 | 46.75 | 64.3 | — | — | — | — | — | — | |
| PConv (MSHNet)Pub’Year=AAAI’25, FLOPs=47.1, Params=15.6, FPS=722026.06 | 92.2 | — | 10.7 | — | — | 67.93 | — | — | — | |
| MSHNetScale loss variant=L3 (Mobius), Gaussian-shaped spatial attention (GP)=true2026.04 | 92.19 | — | 11.46 | — | 67.18 | — | — | — | — | |
| DNA-NetPublication=TIP’222026.01 | 92.18 | 67.54 | 11.77 | — | — | — | — | — | — | |
| L2SKNetPublication=TGRS’252026.01 | 92.18 | 65.67 | 23.99 | — | — | — | — | — | — | |
| DNA-NetType=CNN, Publication=TIP, 20222026.01 | 92.18 | 67.54 | — | 11.77 | — | — | — | — | — | |
| L2SKNetType=CNN, Publication=TGRS, 20252026.01 | 92.18 | 65.67 | — | 23.99 | — | — | — | — | — | |
| ACMMethod Category=Deep Learning Methods2026.03 | 91.92 | 58.84 | 27.59 | — | — | 58.23 | — | — | — | |
| UIU-Net2026.05 | 91.92 | 65.37 | — | 16.85 | — | 64.13 | 78.62 | — | — | |
| MSDANetStrategy=Full2026.05 | 91.92 | 68.08 | 24.05 | — | — | 64.29 | — | — | — | |
| HDNet2026.05 | 91.84 | 65.57 | — | 11.99 | — | 60.35 | 79.2 | — | — | |
| ACMDescription=CNN Full2023.04 | 91.58 | 57.49 | 43.86 | — | — | — | — | — | — | |
| DNANetMethod Category=Deep Learning Methods2026.03 | 91.58 | 66.5 | 18.31 | — | — | 66.13 | — | — | — | |
| L2SKNetMethod Category=Deep Learning Methods2026.03 | 91.58 | 67.38 | 14.78 | — | — | 66.36 | — | — | — | |
| SCTransNet2026.05 | 91.58 | 63.46 | — | 18.28 | — | 62.07 | 77.09 | — | — | |
| ALCLNetStrategy=MCLC2026.05 | 91.58 | 63.11 | 29.8 | — | — | 59.66 | — | — | — | |
| MSHNetScale loss variant=L4 (Var-denominator)2026.04 | 91.5 | — | 17.69 | — | 65.78 | — | — | — | — | |
| UIUNet2025.12 | 91.3 | — | — | — | 65.6 | — | — | — | — | |
| ACMPublish=WACV’212026.05 | 91.25 | 63.39 | 8.961 | — | — | — | — | — | — | |
| ALCLNetStrategy=Full2026.05 | 91.25 | 65.87 | 12.13 | — | — | 64.31 | — | — | — | |
| UIUNetParams (M)=50.542026.06 | 91.24 | — | — | 23.47 | 65.69 | — | 79.31 | — | — | |
| DNA-Net2026.05 | 91.16 | 62.76 | — | 17.61 | — | 58.57 | 77.13 | — | — | |
| AGPC-NetPublication=TAES,’232026.01 | 91.15 | 65.93 | 11.32 | — | — | — | — | — | — | |
| AGPC-NetType=CNN, Publication=TAES, 20232026.01 | 91.15 | 65.93 | — | 11.32 | — | — | — | — | — | |
| DNANetParams (M)=4.692026.06 | 91.13 | — | — | 55.63 | 62.87 | — | 77.2 | — | — | |
| SCAFNet2025.12 | 91.1 | — | — | — | 66.3 | — | — | — | — | |
| ALCNetDescription=CNN Full2023.04 | 90.91 | 62.03 | 42.46 | — | — | — | — | — | — | |
| MPCMType=Local Contrast, Publication=PR, 20162026.01 | 90.91 | 2.3 | — | 525.25 | — | — | — | — | — | |
| UIUNetMethod Category=Deep Learning Methods2026.03 | 90.91 | 65.34 | 15.01 | — | — | 65.55 | — | — | — | |
| DNANet2026.04 | 90.91 | — | 12.24 | — | 66.38 | — | — | — | — | |
| DNANetStrategy=MCLC2026.05 | 90.91 | 61.79 | 32.51 | — | — | 64.32 | — | — | — | |
| MSHNetScale loss variant=L4 (Var-denominator), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 90.82 | — | 24.82 | — | 64.79 | — | — | — | — | |
| DNANetStrategy=Ours2026.05 | 90.82 | 61.38 | 17 | — | — | 58.07 | — | — | — | |
| MSDANetStrategy=Ours2026.05 | 90.82 | 60.61 | 23.76 | — | — | 59.38 | — | — | — | |
| DRPCA-NetParams (M)=1.162026.06 | 90.72 | — | — | 23.09 | 65.23 | — | 78.95 | — | — | |
| ALCNetDescription=CNN Centroid+, Supervision=Centroid Point, Framework=LESPS2023.04 | 90.57 | 44.9 | 84.68 | — | — | — | — | — | — |