Table 2 The ablation experiments on the SVHN dataset.
From: D2S-DiffGAN: a novel image classification model under limited labeled samples
Size of training set | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 |
|---|---|---|---|---|---|
5% | 82.28% | 82.17% | 86.58% | 89.12% | 91.12% |
10% | 85.77% | 86.32% | 89.67% | 91.34% | 93.07% |
15% | 87.91% | 89.25% | 91.54% | 92.84% | 94.13% |
20% | 90.74% | 91.69% | 92.38% | 93.37% | 94.92% |
25% | 91.06% | 92.47% | 93.54% | 94.65% | 95.72% |
