Saliency prediction on MIT1003 (test)
2.9214NSSTranSalNet_Dense
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| TranSalNet_DenseBackbone=DenseNet-161, Pre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.9214 | 0.7547 | 0.9116 | 0.7743 | 0.6279 | 0.7862 | |
| TranSalNet_ResBackbone=ResNet-50, Pre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.8501 | 0.7546 | 0.9093 | 0.7595 | 0.6145 | 0.7779 | |
| MSI-NetPre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.8007 | 0.7454 | 0.9068 | 0.7473 | 0.6081 | 0.8155 | |
| SAM-ResNetBackbone=ResNet-50, Pre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.8001 | 0.7365 | 0.9024 | 0.7466 | 0.6068 | 1.247 | |
| UNISALPre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.7593 | 0.7326 | 0.9026 | 0.734 | 0.5973 | 1.0138 | |
| SAM-VGGBackbone=VGG, Pre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.752 | 0.7256 | 0.9003 | 0.726 | 0.5976 | 1.2195 | |
| DVAPre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.574 | 0.7258 | 0.897 | 0.699 | 0.5663 | 0.7528 | |
| DINet + DCLBackbone=DINet, Optimizer=Adam, Variant=DCL-120-12019.12 | 2.4566 | 0.7611 | 0.8476 | 0.7597 | — | — | |
| DINet + GEMBackbone=DINet, Optimizer=Adam, Baseline=GEM2019.12 | 2.4456 | 0.7571 | 0.8432 | 0.754 | — | — | |
| DINetBackbone=DINet, Optimizer=Adam2019.12 | 2.4406 | 0.757 | 0.8442 | 0.7534 | — | — | |
| ResNet-50 + DCLBackbone=ResNet-50, Optimizer=RMSP, Variant=DCL-oo-12019.12 | 2.4252 | 0.762 | 0.8469 | 0.7658 | — | — | |
| ResNet-50 + DCLBackbone=ResNet-50, Optimizer=Adam, Variant=DCL-oo-12019.12 | 2.4108 | 0.7613 | 0.8442 | 0.7617 | — | — | |
| ResNet-50Backbone=ResNet-50, Optimizer=Adam2019.12 | 2.4064 | 0.7597 | 0.8429 | 0.7618 | — | — | |
| ResNet-50Backbone=ResNet-50, Optimizer=RMSP2019.12 | 2.4047 | 0.7612 | 0.8455 | 0.7595 | — | — | |
| ResNet-50 + GEMBackbone=ResNet-50, Optimizer=RMSP, Baseline=GEM2019.12 | 2.396 | 0.7566 | 0.8412 | 0.75 | — | — | |
| ResNet-50 + GEMBackbone=ResNet-50, Optimizer=Adam, Baseline=GEM2019.12 | 2.3685 | 0.7594 | 0.8427 | 0.7524 | — | — | |
| ML-NetPre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.3329 | 0.7218 | 0.8623 | 0.5979 | 0.496 | 1.3496 | |
| FastSalPre-training=SALICON, Evaluation protocol=10-fold Cross-Validation2021.10 | 2.0078 | 0.706 | 0.8745 | 0.5901 | 0.4783 | 1.0359 |