Few-Shot Image Classification on Omniglot 20-Way
99AccuracyOur GNN
Evaluation Results
| Method | Links | |
|---|---|---|
| Our GNNShots=5-shot2017.11 | 99 | |
| Prototypical NetworksShots=5-shot2017.11 | 98.8 | |
| ConvNet with MemoryShots=5-shot2017.11 | 98.6 | |
| Matching NetworksShots=5-shot2017.11 | 98.5 | |
| N. StatisticianShots=5-shot2017.11 | 98.1 | |
| Our GNNShots=1-shot2017.11 | 97.4 | |
| Siamese NetShots=5-shot2017.11 | 97 | |
| Meta NetworksShots=1-shot2017.11 | 97 | |
| Prototypical NetworksShots=1-shot2017.11 | 95.4 | |
| ConvNet with MemoryShots=1-shot2017.11 | 95 | |
| Res. Pair-WiseShots=1-shot2017.11 | 94.8 | |
| Matching NetworksShots=1-shot2017.11 | 93.8 | |
| N. StatisticianShots=1-shot2017.11 | 93.2 | |
| Siamese NetShots=1-shot2017.11 | 88.2 | |
| PixelsShots=5-shot2017.11 | 42.6 | |
| PixelsShots=1-shot2017.11 | 26.7 |