Phase classification on Pouring (test)
91.82AccuracyTCC
Evaluation Results
| Method | Links | |
|---|---|---|
| TCCBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 91.82 | |
| Supervised LearningBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 91.55 | |
| TCC + TCNBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 91.51 | |
| TCCBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 91.43 | |
| TCC + TCNBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 91.23 | |
| TCC + SaLBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 90.75 | |
| TCC + SaLBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 90.69 | |
| TCNBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 90.39 | |
| TCNBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 90.35 | |
| TCC% of Labels=1.0, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 90.21 | |
| TCC% of Labels=0.5, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 89.43 | |
| TCCBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 89.23 | |
| TCC + SaLBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 89.21 | |
| TCNBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 89.19 | |
| TCC + TCNBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 89.17 | |
| Supervised Learning% of Labels=1.0, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 88.41 | |
| SaLBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 88.02 | |
| SaLBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 87.84 | |
| TCC% of Labels=0.1, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 86.82 | |
| Supervised LearningBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 86.14 | |
| SaLBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 85.68 | |
| TCN% of Labels=1.0, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 84.57 | |
| TCN% of Labels=0.5, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 83.27 | |
| SaL% of Labels=1.0, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 83.19 | |
| SaL% of Labels=0.5, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 80.96 | |
| Supervised Learning% of Labels=0.5, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 77.67 | |
| TCN% of Labels=0.1, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 76.03 | |
| Supervised LearningBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 75.43 | |
| SaL% of Labels=0.1, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 74.5 | |
| Supervised Learning% of Labels=0.1, Backbone=VGG-M, Training Protocol=from scratch2019.04 | 62.01 | |
| ImageNet FeaturesBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 51.13 | |
| Random FeaturesBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=1.02019.04 | 46.08 | |
| ImageNet FeaturesBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 46.06 | |
| Random FeaturesBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.52019.04 | 45.94 | |
| ImageNet FeaturesBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 43.85 | |
| Random FeaturesBackbone=ResNet-50, Pre-trained=ImageNet, Evaluation Protocol=Fine-tuning, % of Labels=0.12019.04 | 42.73 |