Infrared Small Target Detection on NUDT-SIRST
97.45IoUNa-IRSTD
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Na-IRSTD2026.05 | 97.45 | 100 | 1.22 | — | |
| MSDANetStrategy=Full2026.05 | 95.65 | 99.47 | 1.38 | 95.48 | |
| UIUNetStrategy=Full2026.05 | 95.38 | 99.05 | 1.38 | 95.22 | |
| DNANetStrategy=Full2026.05 | 95.35 | 99.26 | 2 | 95.43 | |
| SAISTPublish=CVPR’252026.05 | 95.23 | 99.28 | 1.31 | — | |
| IRMambaPublish=AAAI’252026.05 | 95.18 | 99.26 | 1.31 | — | |
| MSDANetStrategy=Full2026.05 | 94.01 | 98.73 | 1.47 | 94.18 | |
| DNANetStrategy=Full2026.05 | 93.9 | 98.84 | 3.06 | 94.06 | |
| FSGNetMethod Category=Deep Learning Methods2026.03 | 93.78 | 99.26 | 4.89 | 94.14 | |
| L2SKNetMethod Category=Deep Learning Methods2026.03 | 93.52 | 97.57 | 5.29 | 92.86 | |
| DNANetPublish=TIP’222026.05 | 92.09 | 99.53 | 2.35 | — | |
| DNANetMethod Category=Deep Learning Methods2026.03 | 91.96 | 98.94 | 9.28 | 92.88 | |
| MLCLNetStrategy=Full2026.05 | 91.25 | 97.67 | 10.09 | 91.54 | |
| UIUNetMethod Category=Deep Learning Methods2026.03 | 90.43 | 98.94 | 7.72 | 89.87 | |
| UIUNetPublish=TIP’232026.05 | 90.17 | 99.29 | 2.39 | — | |
| ALCLNetStrategy=Full2026.05 | 89.79 | 98.84 | 6.69 | 89.95 | |
| RPCANetPublish=WACV’242026.05 | 89.03 | 97.14 | 28.7 | — | |
| ISNetPublish=CVPR’222026.05 | 88.92 | 99.12 | 4.21 | — | |
| AGPCNetMethod Category=Deep Learning Methods2026.03 | 88.87 | 97.2 | 10.02 | 90.64 | |
| AGPCNetPublish=TAES’232026.05 | 88.71 | 97.57 | 7.54 | — | |
| RPCANetMethod Category=Deep Learning Methods2026.03 | 87.79 | 95.98 | 32.03 | 89.8 | |
| DNA-NetDescription=CNN Full2023.04 | 87.42 | 98.31 | 24.5 | — | |
| MSHNetScale loss variant=L3 (Mobius)2026.04 | 82.9 | 97.5 | 12.2 | — | |
| MSHNetScale loss variant=L1 (Diff-based), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 82.86 | 97.4 | 20.89 | — | |
| MSHNetScale loss variant=L3 (Mobius), Gaussian-shaped spatial attention (GP)=true2026.04 | 82.74 | 97.53 | 11.74 | — | |
| MSHNetScale loss variant=L1 (Diff-based), Gaussian-shaped spatial attention (GP)=true2026.04 | 82.55 | 97.42 | 16.16 | — | |
| MSHNetScale loss variant=L3 (Mobius), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 81.44 | 96.3 | 10.68 | — | |
| MSHNetScale loss variant=L1 (Diff-based), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 81.38 | 95.75 | 16.24 | — | |
| MSHNetScale loss variant=L1 (Diff-based)2026.04 | 80.77 | 97.33 | 12.87 | — | |
| SCTransNet2026.04 | 80.69 | 95.71 | 16.52 | — | |
| DNANetStrategy=Ours2026.05 | 79.99 | 97.67 | 13.65 | 87.01 | |
| MSHNetScale loss variant=L4 (Var-denominator)2026.04 | 79.95 | 96.28 | 12.55 | — | |
| ALCLNetStrategy=Ours2026.05 | 79.88 | 96.8 | 17.27 | 85.5 | |
| REEMDescription=SCR-Aware Reweighting2026.06 | 79.86 | 97.52 | 11.21 | — | |
| UIUNetStrategy=Ours2026.05 | 79.84 | 98.55 | 6.77 | 85.19 | |
| UIUNetStrategy=Ours2026.05 | 79.33 | 97.67 | 15.95 | 84.22 | |
| DNANetStrategy=Ours2026.05 | 79.32 | 98.26 | 10.21 | 87.2 | |
| MSHNetScale loss variant=L2 (Var-based)2026.04 | 79.23 | 98.84 | 11.64 | — | |
| MSDANetStrategy=Ours2026.05 | 79.08 | 98.84 | 13.19 | 86.05 | |
| MLCLNetStrategy=Ours2026.05 | 78.68 | 97.67 | 12.79 | 85.7 | |
| MSHNetScale loss variant=L2 (Var-based), Gaussian-shaped spatial attention (GP)=true2026.04 | 78.45 | 96.46 | 10.01 | — | |
| DNANetDescription=Deep Learning2026.06 | 78.28 | 93.33 | 15.01 | — | |
| MSHNetScale loss variant=L2 (Var-based), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 78.21 | 97.6 | 21.25 | — | |
| UIU-Net2026.04 | 78.09 | 96.77 | 12.54 | — | |
| MSHNetScale loss variant=L4 (Var-denominator), Gaussian-shaped spatial attention (GP)=true, Learnable rotated pinwheel mask=true2026.04 | 78.01 | 97.4 | 27.5 | — | |
| RDIANMethod Category=Deep Learning Methods2026.03 | 77.77 | 95.66 | 26.36 | 79.46 | |
| MTU-Net2026.04 | 77.54 | 93.97 | 46.95 | — | |
| MSHNetScale loss variant=L2 (Var-based), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 77.39 | 97.2 | 22.27 | — | |
| MSHNetMethod Category=Deep Learning Methods2026.03 | 76.95 | 95.77 | 20.52 | 79.72 | |
| MSHNetScale loss variant=L4 (Var-denominator), Gaussian-shaped spatial attention (GP)=true2026.04 | 76.79 | 93.5 | 22.93 | — | |
| DNANet2026.04 | 76.47 | 97.31 | 14.13 | — | |
| MSHNetScale loss variant=L3 (Mobius), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 76.45 | 97.1 | 22.24 | — | |
| UIU-NetDescription=CNN Full2023.04 | 75.91 | 96.83 | 18.61 | — | |
| MSHNetDescription=Deep Learning2026.06 | 74.52 | 95.37 | 29 | — | |
| MSHNetScale loss variant=L4 (Var-denominator), Location regularizer=omitted, Gaussian-shaped spatial attention (GP)=true2026.04 | 73.09 | 97 | 38.68 | — | |
| ALCNetDescription=Deep Learning2026.06 | 72.89 | 96.19 | 30.4 | — | |
| ISNetDescription=CNN Full2023.04 | 71.27 | 96.93 | 96.84 | — | |
| DNANetStrategy=MCLC2026.05 | 70.86 | 96.19 | 25 | 72.82 | |
| UIUNetStrategy=PAL2026.05 | 70.81 | 98.2 | 13.42 | 72.43 | |
| DNANetStrategy=PAL2026.05 | 70.7 | 97.78 | 24.57 | 71.34 | |
| UIUNetStrategy=MCLC2026.05 | 70.2 | 94.71 | 21.46 | 72.22 | |
| MSDANetStrategy=MCLC2026.05 | 69.67 | 94.92 | 26.52 | 71.61 | |
| ACMPublish=WACV’212026.05 | 69.26 | 96.26 | 10.27 | — | |
| ALCLNetStrategy=MCLC2026.05 | 68.89 | 95.45 | 39.94 | 70.84 | |
| ISNetDescription=Deep Learning2026.06 | 67.86 | 92.59 | 34.65 | — | |
| ALCLNetStrategy=PAL2026.05 | 67.4 | 96.61 | 23 | 68.89 | |
| ACM2026.04 | 64.4 | 93.12 | 55.22 | — | |
| DNA-NetDescription=CNN Centroid+, Supervision=Centroid Point, Framework=LESPS2023.04 | 62.73 | 93.27 | 21.81 | — | |
| ALCNetDescription=CNN Full2023.04 | 61.78 | 91.32 | 36.36 | — | |
| ACMMethod Category=Deep Learning Methods2026.03 | 60.63 | 91.75 | 20.06 | 60.94 | |
| MSDANetStrategy=PAL2026.05 | 59.46 | 93.54 | 16.13 | 63.76 | |
| DNA-NetDescription=CNN Coarse+, Supervision=Coarse Point, Framework=LESPS2023.04 | 58.37 | 93.76 | 28.01 | — | |
| ACMDescription=CNN Full2023.04 | 57.42 | 91.75 | 39.73 | — | |
| MDvsFADescription=CNN Full2023.04 | 45.38 | 86.03 | 200.71 | — | |
| ALCNetDescription=CNN Coarse+, Supervision=Coarse Point, Framework=LESPS2023.04 | 44.14 | 92.8 | 32.1 | — | |
| ACMDescription=CNN Centroid+, Supervision=Centroid Point, Framework=LESPS2023.04 | 42.09 | 91.11 | 38.24 | — | |
| ALCNetDescription=CNN Centroid+, Supervision=Centroid Point, Framework=LESPS2023.04 | 41.58 | 92.28 | 67.01 | — | |
| ACMDescription=CNN Coarse+, Supervision=Coarse Point, Framework=LESPS2023.04 | 40.64 | 81.11 | 49.45 | — | |
| UIUNetStrategy=LESPS2026.05 | 39.63 | 68.04 | 20.04 | 46.56 | |
| DNANetStrategy=LESPS2026.05 | 36.98 | 66.14 | 31.97 | 46.46 | |
| ALCLNetStrategy=LESPS2026.05 | 36.76 | 73.02 | 39.02 | 44.9 | |
| MSDANetStrategy=LESPS2026.05 | 36.34 | 73.76 | 19.4 | 45.45 | |
| MDvsFADescription=Deep Learning2026.06 | 35.86 | 85.22 | 95.37 | — | |
| RIPTDescription=Low Rank2026.06 | 29.44 | 91.85 | 344.3 | — | |
| RIPTMethod Category=Low-rank and Sparse Decomposition Methods2026.03 | 29.17 | 91.85 | 344.3 | 36.12 | |
| IPIMethod Category=Low-rank and Sparse Decomposition Methods2026.03 | 28.63 | 74.49 | 41.23 | 38.18 | |
| PSTNNMethod Category=Low-rank and Sparse Decomposition Methods2026.03 | 27.72 | 66.13 | 44.17 | 39.8 | |
| PSTNNPublish=RS’192026.05 | 27.72 | 66.13 | 44.17 | — | |
| NOLCMethod Category=Low-rank and Sparse Decomposition Methods2026.03 | 23.87 | 85.47 | 58.2 | 34.9 | |
| Top-HatDescription=Filtering2023.04 | 20.72 | 78.41 | 166.7 | — | |
| Top-Hat (Morphological)Description=Filtering2026.06 | 20.72 | 78.41 | 166.7 | — | |
| IPIDescription=Low Rank2023.04 | 17.76 | 74.49 | 41.23 | — | |
| IPIDescription=Low Rank2026.06 | 17.76 | 74.49 | 41.23 | — | |
| RLCMDescription=Local Contrast2023.04 | 15.14 | 66.35 | 163 | — | |
| PSTNNDescription=Low Rank2023.04 | 14.85 | 66.13 | 44.17 | — | |
| PSTNNDescription=Low Rank2026.06 | 14.85 | 66.13 | 44.17 | — | |
| FKRWMethod Category=Background Suppression Methods2026.03 | 12.67 | 79.51 | 67.13 | 21.73 | |
| NRAMPublish=RS’182026.05 | 12.08 | 72.58 | 84.77 | — | |
| NWMTHMethod Category=Background Suppression Methods2026.03 | 11.72 | 52.71 | 46.81 | 10.26 | |
| TLLCMMethod Category=Background Suppression Methods2026.03 | 11.08 | 77.97 | 67.26 | 23.82 |