Salient Object Detection on HKU-IS 18 (test)
93.6F-betaPoolNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PoolNetBackbone=VGG-16, Training #Images=5,000, Training Dataset=MSRA-B + HKU-IS, joint training with edge detection=false2019.04 | 93.6 | — | 0.032 | |
| PoolNetBackbone=ResNet-50, Training #Images=10,553, Training Dataset=DUTS-TR, joint training with edge detection=true2019.04 | 93.5 | — | 0.03 | |
| GearNetBackbone=VGG-16, Training #Images=5,000, Training Dataset=MSRA-B + HKU-IS2019.04 | 93.4 | — | 0.034 | |
| PoolNetBackbone=ResNet-50, Training #Images=10,553, Training Dataset=DUTS-TR, joint training with edge detection=false2019.04 | 93.4 | — | 0.032 | |
| PoolNetBackbone=VGG-16, Training #Images=10,553, Training Dataset=DUTS-TR, joint training with edge detection=true2019.04 | 93.1 | — | 0.033 | |
| PoolNetBackbone=VGG-16, Training #Images=10,553, Training Dataset=DUTS-TR, joint training with edge detection=false2019.04 | 92.8 | — | 0.035 | |
| PiCANet-RCBackbone=ResNet50, Post-processing (CRF)=true2017.08 | 92.7 | 89 | 0.031 | |
| PiCANet-CBackbone=VGG-16, Post-processing (CRF)=true2017.08 | 92.5 | 88.9 | 0.031 | |
| PiCANetBackbone=VGG-16, Post-processing (CRF)=false2017.08 | 92.1 | 84.7 | 0.042 | |
| PiCANetBackbone=VGG-16, Training #Images=10,553, Training Dataset=DUTS-TR2019.04 | 92.1 | — | 0.042 | |
| PiCANetBackbone=ResNet-50, Training #Images=10,553, Training Dataset=DUTS-TR2019.04 | 92 | — | 0.044 | |
| PiCANet-RBackbone=ResNet50, Post-processing (CRF)=false2017.08 | 91.9 | 84 | 0.043 | |
| PAGRBackbone=VGG-16, Training #Images=10,553, Training Dataset=DUTS-TR2019.04 | 91.9 | — | 0.047 | |
| DSSBackbone=VGG-16, Post-processing (CRF)=true2017.08 | 91.1 | 86.6 | 0.04 | |
| DGRLBackbone=ResNet-50, Training #Images=10,553, Training Dataset=DUTS-TR2019.04 | 91 | — | 0.036 | |
| PoolNetBackbone=VGG-16, Training #Images=2,500, Training Dataset=MSRA-B, joint training with edge detection=false2019.04 | 90.8 | — | 0.044 | |
| MSRBackbone=VGG-16, Training #Images=5,000, Training Dataset=MSRA-B + HKU-IS2019.04 | 90.7 | — | 0.043 | |
| SRMBackbone=ResNet50, Post-processing (CRF)=false2017.08 | 90.6 | 83.6 | 0.046 | |
| SRMBackbone=ResNet-50, Training #Images=10,553, Training Dataset=DUTS-TR2019.04 | 90.6 | — | 0.046 | |
| DHSBackbone=VGG-16, Post-processing (CRF)=false2017.08 | 90.2 | 80.6 | 0.054 | |
| NLDFBackbone=VGG-16, Post-processing (CRF)=false2017.08 | 90.2 | 83.8 | 0.048 | |
| NLDFBackbone=VGG-16, Training #Images=3,000, Training Dataset=MSRA-B2019.04 | 90.2 | — | 0.048 | |
| DSSBackbone=VGG-16, Training #Images=2,500, Training Dataset=MSRA-B2019.04 | 90 | — | 0.05 | |
| RFCNBackbone=VGG-16, Post-processing (CRF)=false2017.08 | 89.8 | 71.8 | 0.08 | |
| AmuletBackbone=VGG-16, Post-processing (CRF)=false2017.08 | 89.6 | 81.3 | 0.052 | |
| RFCNBackbone=VGG-16, Training #Images=10,000, Training Dataset=MSRA10K2019.04 | 89.5 | — | 0.079 | |
| DCLBackbone=VGG-16, Training #Images=2,500, Training Dataset=MSRA-B2019.04 | 89.3 | — | 0.063 | |
| DHSBackbone=VGG-16, Training #Images=9,500, Training Dataset=MSRA10K + DUT-OMRON2019.04 | 89.2 | — | 0.052 | |
| AmuletBackbone=VGG-16, Training #Images=10,000, Training Dataset=MSRA10K2019.04 | 88.9 | — | 0.052 | |
| UCFBackbone=VGG-16, Training #Images=10,000, Training Dataset=MSRA10K2019.04 | 88.8 | — | 0.073 | |
| UCFBackbone=VGG-16, Post-processing (CRF)=false2017.08 | 88.6 | 75.1 | 0.074 | |
| DCLBackbone=VGG-16, Post-processing (CRF)=true2017.08 | 88.5 | 73.6 | 0.137 |