Saliency Prediction on DHF1K
15.5Model Size (MB)UNISAL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| UNISALMNet V2 encoder=true, Separable convolutions=true, CGRU at bottleneck=true, Bilinear upsampling=true2020.03 | 15.5 | 0.009 | |
| UNISALMNet V2 encoder=true, Separable convolutions=false, CGRU at bottleneck=false, Bilinear upsampling=false2020.03 | 59.4 | 0.017 | |
| TASED-Net2020.03 | 82 | 0.06 | |
| DVA2020.03 | 96 | 0.1 | |
| Deep-Net2020.03 | 103 | 0.08 | |
| SALICON2020.03 | 117 | 0.5 | |
| SalGAN2020.03 | 130 | 0.02 | |
| ACLNet2020.03 | 250 | 0.02 | |
| Two-stream2020.03 | 315 | 20 | |
| SalEMA2020.03 | 364 | 0.01 | |
| STRA-Net2020.03 | 641 | 0.02 | |
| Shallow-Net2020.03 | 2,500 | 0.1 |