Image Classification on MNIST to SVHN standard (test)
97.16AccuracySupTgt
Evaluation Results
| Method | Links | |
|---|---|---|
| SupTgtTraining Protocol=Train on Target, Augmentation Type=Intensity Augmentation2017.06 | 97.16 | |
| Mean TeacherTraining Protocol=Unsupervised Domain Adaptation, Augmentation Type=Intensity Augmentation, Note=with class balance loss2017.06 | 97 | |
| SupTgtTraining Protocol=Train on Target, Augmentation Type=TFA2017.06 | 96.65 | |
| SupTgtTraining Protocol=Train on Target, Augmentation Type=TF2017.06 | 96.19 | |
| SupTgtTraining Protocol=Train on Target, Augmentation Type=minimal2017.06 | 95.7 | |
| DIRT-TBackbone=Instance Normalization2019.09 | 76.5 | |
| VADABackbone=Instance Normalization2019.09 | 73.3 | |
| Π-Model2019.09 | 71.4 | |
| SupSrcTraining Protocol=Train on Source, Augmentation Type=Intensity Augmentation2017.06 | 61.99 | |
| Proposed MethodTask=Rotation2019.09 | 61.3 | |
| SBADA-GANTraining Protocol=Unsupervised Domain Adaptation2017.06 | 61.08 | |
| DIRT-T2019.09 | 54.5 | |
| ATTTraining Protocol=Unsupervised Domain Adaptation2017.06 | 52.8 | |
| ATT2019.09 | 52.8 | |
| VADA2019.09 | 47.5 | |
| Source only (VADA & DIRT-T)mode=source only, Backbone=Instance Normalization2019.09 | 40.9 | |
| KNN-Ad2019.09 | 40.3 | |
| DRCN2019.09 | 40.1 | |
| DCRNTraining Protocol=Unsupervised Domain Adaptation2017.06 | 40.05 | |
| Mean TeacherTraining Protocol=Unsupervised Domain Adaptation, Augmentation Type=TFA, Confidence Thresholding=true, Note=with class balance loss2017.06 | 37.49 | |
| G2ATraining Protocol=Unsupervised Domain Adaptation2017.06 | 36.4 | |
| DANN2019.09 | 35.7 | |
| RevGradTraining Protocol=Unsupervised Domain Adaptation2017.06 | 35.67 | |
| Mean TeacherTraining Protocol=Unsupervised Domain Adaptation, Augmentation Type=TF, Confidence Thresholding=true, Note=with class balance loss2017.06 | 34.15 | |
| Mean TeacherTraining Protocol=Unsupervised Domain Adaptation, Augmentation Type=minimal, Confidence Thresholding=true, Note=with class balance loss2017.06 | 33.87 | |
| Source only (Our method)mode=source only2019.09 | 30.5 | |
| SupSrcTraining Protocol=Train on Source, Augmentation Type=TFA2017.06 | 28.69 | |
| Source only (VADA & DIRT-T)mode=source only2019.09 | 27.9 | |
| SupSrcTraining Protocol=Train on Source, Augmentation Type=minimal2017.06 | 25.44 | |
| SupSrcTraining Protocol=Train on Source, Augmentation Type=TF2017.06 | 24.86 | |
| Mean TeacherTraining Protocol=Unsupervised Domain Adaptation, Augmentation Type=TF, Note=with class balance loss2017.06 | 13.96 |