Salient Object Detection on DUTS (test)
0.015M (MAE)S3ODNet
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| S3ODNetData=S3OD + SOD2025.10 | 0.015 | — | — | — | — | — | — | 0.972 | 0.954 | — | — | 0.949 | |
| BED-SAM2Size=512, TR=[1,2,3]2026.05 | 0.017 | 0.944 | 0.946 | — | — | — | — | 0.965 | — | — | — | — | |
| BED-SAM2Size=1024, TR=[1,2]2026.05 | 0.017 | 0.943 | 0.945 | — | — | — | — | 0.963 | — | — | — | — | |
| BED-SAM2Size=1024, TR=[1,2,3]2026.05 | 0.017 | 0.944 | 0.947 | — | — | — | — | 0.964 | — | — | — | — | |
| BiRefNetTraining sets (TR)=1, 22024.01 | 0.018 | 0.938 | 0.935 | 0.96 | — | — | — | — | — | — | — | — | |
| BiRefNetTraining sets (TR)=1, 32024.01 | 0.018 | 0.942 | 0.942 | 0.961 | — | — | — | — | — | — | — | — | |
| BiRefNetTraining sets (TR)=1, 2, 32024.01 | 0.018 | 0.944 | 0.943 | 0.962 | — | — | — | — | — | — | — | — | |
| BiRefNetData=SOD2025.10 | 0.018 | — | — | — | — | — | — | 0.962 | 0.943 | — | — | 0.944 | |
| S3ODNetData=SOD2025.10 | 0.018 | — | — | — | — | — | — | 0.966 | 0.951 | — | — | 0.939 | |
| BiRefNetSize=1024, TR=[1,2]2026.05 | 0.018 | 0.938 | 0.935 | — | — | — | — | 0.96 | — | — | — | — | |
| BiRefNetSize=1024, TR=[1,3]2026.05 | 0.018 | 0.942 | 0.942 | — | — | — | — | 0.961 | — | — | — | — | |
| BiRefNetSize=1024, TR=[1,2,3]2026.05 | 0.018 | 0.944 | 0.943 | — | — | — | — | 0.962 | — | — | — | — | |
| BED-SAM2Size=512, TR=12026.05 | 0.018 | 0.943 | 0.942 | — | — | — | — | 0.963 | — | — | — | — | |
| BED-SAM2Size=512, TR=[1,2]2026.05 | 0.018 | 0.943 | 0.943 | — | — | — | — | 0.963 | — | — | — | — | |
| BED-SAM2Size=1024, TR=12026.05 | 0.018 | 0.941 | 0.94 | — | — | — | — | 0.96 | — | — | — | — | |
| BED-SAM2Size=1024, TR=[1,3]2026.05 | 0.018 | 0.942 | 0.944 | — | — | — | — | 0.963 | — | — | — | — | |
| BiRefNetTraining sets (TR)=12024.01 | 0.019 | 0.939 | 0.937 | 0.958 | — | — | — | — | — | — | — | — | |
| BiRefNetSize=1024, TR=12026.05 | 0.019 | 0.939 | 0.937 | — | — | — | — | 0.958 | — | — | — | — | |
| BED-SAM2Size=512, TR=[1,3]2026.05 | 0.019 | 0.944 | 0.942 | — | — | — | — | 0.961 | — | — | — | — | |
| BiRefNetTraining sets (TR)=2, 32024.01 | 0.02 | 0.933 | 0.928 | 0.954 | — | — | — | — | — | — | — | — | |
| SAM2-UNet2024.08 | 0.02 | — | — | — | — | — | — | 0.959 | — | — | — | 0.934 | |
| S3ODNetTraining Data=S3OD2025.10 | 0.02 | — | — | 0.962 | — | 0.938 | — | — | 0.937 | — | — | — | |
| BiRefNetSize=1024, TR=[2,3]2026.05 | 0.02 | 0.933 | 0.928 | — | — | — | — | 0.954 | — | — | — | — | |
| SAM2-UNetSize=352, TR=12026.05 | 0.02 | 0.934 | — | — | — | — | — | 0.959 | — | — | — | — | |
| BED-SAM2Size=352, TR=[1,2]2026.05 | 0.02 | 0.939 | 0.938 | — | — | — | — | 0.959 | — | — | — | — | |
| BED-SAM2Size=352, TR=[1,3]2026.05 | 0.02 | 0.94 | 0.939 | — | — | — | — | 0.961 | — | — | — | — | |
| BED-SAM2Size=352, TR=[1,2,3]2026.05 | 0.02 | 0.941 | 0.941 | — | — | — | — | 0.962 | — | — | — | — | |
| BED-SAM2Size=352, TR=12026.05 | 0.021 | 0.937 | 0.933 | — | — | — | — | 0.959 | — | — | — | — | |
| TE7#Params=66.27M, GFLOPS=18.862021.12 | 0.022 | 0.919 | — | — | — | — | — | — | 0.932 | 0.904 | — | — | |
| BED-SAM2Size=512, TR=[2,3]2026.05 | 0.022 | 0.934 | 0.934 | — | — | — | — | 0.954 | — | — | — | — | |
| InSPyReNetbackbone=SwinB, training_datasets=DUTS-TR2022.09 | 0.024 | — | — | — | — | 0.931 | — | — | 0.927 | — | — | — | |
| InSPyReNetBackbone=SwinB2022.09 | 0.024 | — | — | — | — | 0.931 | — | — | 0.927 | — | — | — | |
| M3Net-SArchitecture=Transformer, Backbone=SwinTransformer2023.09 | 0.024 | 0.927 | — | 0.96 | — | — | — | — | 0.902 | — | — | — | |
| InSpyreNetData=SOD2025.10 | 0.024 | — | — | — | — | — | — | 0.956 | 0.932 | — | — | 0.936 | |
| BED-SAM2Size=352, TR=[2,3]2026.05 | 0.024 | 0.93 | 0.927 | — | — | — | — | 0.95 | — | — | — | — | |
| BED-SAM2Size=1024, TR=[2,3]2026.05 | 0.024 | 0.926 | 0.927 | — | — | — | — | 0.949 | — | — | — | — | |
| U2Mamba2026.06 | 0.024 | — | — | — | — | — | — | — | 0.904 | — | — | — | |
| TE6#Params=43.47M, GFLOPS=14.842021.12 | 0.025 | 0.915 | — | — | — | — | — | — | 0.928 | 0.896 | — | — | |
| Mao et al.Backbone=SwinB2022.09 | 0.025 | — | — | — | — | 0.917 | — | — | 0.911 | — | — | — | |
| BBRF2023.05 | 0.025 | — | — | 0.927 | — | 0.908 | — | — | 0.916 | — | — | — | |
| ICON-SArchitecture=Transformer, Backbone=SwinTransformer2023.09 | 0.025 | 0.917 | — | 0.96 | — | — | — | — | 0.886 | — | — | — | |
| BBRFArchitecture=Transformer2023.09 | 0.025 | 0.909 | — | 0.952 | — | — | — | — | 0.886 | — | — | — | |
| Dual-SAMzero-shot=false2024.04 | 0.025 | — | 0.885 | — | — | — | — | — | — | — | — | — | |
| ICON-SSize=352, TR=12026.05 | 0.025 | 0.917 | 0.886 | — | — | — | — | 0.954 | — | — | — | — | |
| TE5#Params=31.3M, GFLOPS=11.522021.12 | 0.026 | 0.909 | — | — | — | — | — | — | 0.918 | 0.883 | — | — | |
| SelfReformer-fullTransformer-based=true, Upsampling Strategy=Pixel Shuffle2022.05 | 0.026 | — | 0.916 | 0.92 | — | 0.911 | — | — | — | — | — | — | |
| SelfReformer2023.05 | 0.026 | — | — | 0.92 | — | 0.911 | — | — | 0.916 | — | — | — | |
| EVPv1Backbone=SegFormer2023.05 | 0.026 | — | — | 0.947 | — | 0.913 | — | — | 0.923 | — | — | — | |
| EVPv2Backbone=ViT-SAM2023.05 | 0.026 | — | — | 0.944 | — | 0.915 | — | — | 0.921 | — | — | — | |
| ARNetSource=20252025.12 | 0.026 | — | 0.883 | 0.947 | — | 0.913 | — | — | — | — | — | — | |
| PGNetTraining sets (TR)=12024.01 | 0.027 | 0.911 | 0.917 | 0.922 | — | — | — | — | — | — | — | — | |
| PGNetTrained on=DUTS-TR, Backbone=ResNet-502022.04 | 0.027 | — | — | 0.922 | — | 0.911 | — | — | 0.917 | — | — | — | |
| PGNetbackbone=SwinB + ResNet18, training_datasets=DUTS-TR2022.09 | 0.027 | — | — | — | — | 0.911 | — | — | 0.903 | — | — | — | |
| EVPv2Backbone=SegFormer2023.05 | 0.027 | — | — | 0.948 | — | 0.915 | — | — | 0.923 | — | — | — | |
| SRformerArchitecture=Transformer2023.09 | 0.027 | 0.91 | — | 0.952 | — | — | — | — | 0.872 | — | — | — | |
| GPONetSource=PR242025.12 | 0.027 | — | — | 0.936 | — | 0.919 | — | — | — | — | — | — | |
| PGNetSize=1024, TR=12026.05 | 0.027 | 0.911 | 0.917 | — | — | — | — | 0.922 | — | — | — | — | |
| PGNetTraining sets (TR)=1, 22024.01 | 0.028 | 0.912 | 0.919 | 0.925 | — | — | — | — | — | — | — | — | |
| PGNetTrained on=DUTS-TR and HRSOD-TR, Backbone=ResNet-502022.04 | 0.028 | — | — | 0.925 | — | 0.912 | — | — | 0.919 | — | — | — | |
| TE3#Params=14.02M, GFLOPS=6.262021.12 | 0.028 | 0.898 | — | — | — | — | — | — | 0.909 | 0.873 | — | — | |
| TE4#Params=20.71M, GFLOPS=8.642021.12 | 0.028 | 0.902 | — | — | — | — | — | — | 0.915 | 0.879 | — | — | |
| VAN-B2Backbone=VAN-B22022.02 | 0.028 | — | — | — | — | — | — | — | 0.919 | — | — | — | |
| PA-KRNbackbone=SwinB, training_datasets=DUTS-TR, re-implemented=true2022.09 | 0.028 | — | — | — | — | 0.913 | — | — | 0.906 | — | — | — | |
| PGNetbackbone=SwinB + ResNet18, training_datasets=DUTS-TR, HRSOD-TR2022.09 | 0.028 | — | — | — | — | 0.912 | — | — | 0.905 | — | — | — | |
| InSPyReNetBackbone=Res2Net502022.09 | 0.028 | — | — | — | — | 0.904 | — | — | 0.892 | — | — | — | |
| PA-KRNBackbone=SwinB, Re-implementation=true2022.09 | 0.028 | — | — | — | — | 0.913 | — | — | 0.906 | — | — | — | |
| MENet2024.08 | 0.028 | — | — | — | — | — | — | 0.937 | — | — | — | 0.905 | |
| DPNetSource=TIP222025.12 | 0.028 | — | — | 0.943 | — | 0.912 | — | — | — | — | — | — | |
| MENetSource=CVPR232025.12 | 0.028 | — | — | 0.938 | — | 0.905 | — | — | — | — | — | — | |
| MENetLabel=F2026.03 | 0.028 | 0.905 | — | — | — | — | — | — | — | — | — | — | |
| MENetSize=352, TR=12026.05 | 0.028 | 0.905 | 0.912 | — | — | — | — | 0.937 | — | — | — | — | |
| PGNetSize=1024, TR=[1,2]2026.05 | 0.028 | 0.912 | 0.919 | — | — | — | — | 0.925 | — | — | — | — | |
| MENET2026.06 | 0.028 | — | — | — | — | — | — | — | 0.895 | — | — | — | |
| C4Net-C5Backbone=ResNet-1012021.10 | 0.029 | — | — | — | 0.886 | — | — | 0.937 | — | — | — | — | |
| Ours2021.12 | 0.029 | — | 0.875 | — | — | 0.908 | — | 0.942 | — | — | — | — | |
| SelfReformer-BITransformer-based=true, Upsampling Strategy=Bilinear Interpolation2022.05 | 0.029 | — | 0.905 | 0.919 | — | 0.904 | — | — | — | — | — | — | |
| MINetbackbone=SwinB, training_datasets=DUTS-TR, re-implemented=true2022.09 | 0.029 | — | — | — | — | 0.906 | — | — | 0.893 | — | — | — | |
| MINetBackbone=SwinB, Re-implementation=true2022.09 | 0.029 | — | — | — | — | 0.906 | — | — | 0.893 | — | — | — | |
| SAM-Adapter2024.04 | 0.029 | — | 0.838 | — | — | — | — | — | — | — | — | — | |
| VST-S++Label=F2026.03 | 0.029 | 0.909 | — | — | — | — | — | — | — | — | — | — | |
| VST-S++2026.06 | 0.029 | — | — | — | — | — | — | — | 0.897 | — | — | — | |
| C4Net-C4Backbone=ResNet-502021.10 | 0.03 | — | — | — | 0.875 | — | — | 0.929 | — | — | — | — | |
| TE2#Params=11.09M, GFLOPS=5.22021.12 | 0.03 | 0.891 | — | — | — | — | — | — | 0.9 | 0.866 | — | — | |
| VAN-B1Backbone=VAN-B12022.02 | 0.03 | — | — | — | — | — | — | — | 0.912 | — | — | — | |
| EVPv2Backbone=ViT-MAE2023.05 | 0.03 | — | — | 0.925 | — | 0.904 | — | — | 0.913 | — | — | — | |
| CFNetSource=PR222025.12 | 0.03 | — | — | 0.941 | — | 0.909 | — | — | — | — | — | — | |
| FCL-CODLabel=B2026.03 | 0.03 | 0.907 | — | — | — | — | — | — | — | — | — | — | |
| DHQTraining sets (TR)=1, 22024.01 | 0.031 | 0.894 | 0.9 | 0.919 | — | — | — | — | — | — | — | — | |
| DHQSODTrained on=DUTS-TR and HRSOD-TR, Backbone=ResNet-502022.04 | 0.031 | — | — | 0.919 | — | 0.894 | — | — | 0.9 | — | — | — | |
| Tang et al.backbone=ResNet50, training_datasets=DUTS-TR, HRSOD-TR2022.09 | 0.031 | — | — | — | — | 0.895 | — | — | 0.888 | — | — | — | |
| DHQData=SOD2025.10 | 0.031 | — | — | — | — | — | — | 0.919 | 0.894 | — | — | 0.9 | |
| SCNet2021.12 | 0.032 | — | 0.87 | — | — | 0.902 | — | 0.936 | — | — | — | — | |
| PVT-SBackbone=PVT-S2022.02 | 0.032 | — | — | — | — | — | — | — | 0.9 | — | — | — | |
| PAKRNTransformer-based=false2022.05 | 0.032 | — | 0.906 | 0.916 | — | 0.9 | — | — | — | — | — | — | |
| LDFBackbone=Res2Net50, Re-implementation=true2022.09 | 0.032 | — | — | — | — | 0.897 | — | — | 0.885 | — | — | — | |
| TE1#Params=9.96M, GFLOPS=4.282021.12 | 0.033 | 0.885 | — | — | — | — | — | — | 0.893 | 0.856 | — | — | |
| LDFTransformer-based=false2022.05 | 0.033 | — | 0.897 | 0.909 | — | 0.892 | — | — | — | — | — | — | |
| F³Netbackbone=SwinB, training_datasets=DUTS-TR, re-implemented=true2022.09 | 0.033 | — | — | — | — | 0.902 | — | — | 0.895 | — | — | — | |
| F³NetBackbone=Res2Net50, Re-implementation=true2022.09 | 0.033 | — | — | — | — | 0.892 | — | — | 0.876 | — | — | — | |
| F³NetBackbone=SwinB, Re-implementation=true2022.09 | 0.033 | — | — | — | — | 0.902 | — | — | 0.895 | — | — | — |