Image Classification on COVIDx CRX-2 (1000 samples)
93.97Mean AccuracyLS+WRN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LS+WRNInit.=TF, Parameterization=Equivariant2021.07 | 93.97 | 0.34 | |
| LS+WRNInit.=TF, Parameterization=Canonical2021.07 | 93.68 | 0.35 | |
| S+WRNInit.=TF, Parameterization=Fixed2021.07 | 91.9 | 0.54 | |
| LS+WRNInit.=Rand, Parameterization=Canonical2021.07 | 91.65 | 0.55 | |
| LS+WRNInit.=Rand, Parameterization=Equivariant2021.07 | 91.4 | 0.67 | |
| ResNet-502021.07 | 90.86 | 0.57 | |
| S+WRNInit.=Rand, Parameterization=Fixed2021.07 | 90.6 | 0.5 | |
| LS+LLInit.=TF, Parameterization=Canonical2021.07 | 89.9 | 0.4 | |
| LS+LLInit.=Rand, Parameterization=Canonical2021.07 | 89.7 | 0.65 | |
| S+LLInit.=TF, Parameterization=Fixed2021.07 | 89.23 | 0.69 | |
| LS+LLInit.=TF, Parameterization=Equivariant2021.07 | 89.15 | 0.6 | |
| WRN-162021.07 | 88.82 | 1.64 | |
| LS+LLInit.=Rand, Parameterization=Equivariant2021.07 | 87.98 | 0.55 | |
| S+LLInit.=Rand, Parameterization=Fixed2021.07 | 86.8 | 0.41 |