Table 2 Few-shot specific implementation information including model parameters, image information, and computing device used for the images shown in Fig. 2.
From: Rapid and flexible segmentation of electron microscopy data using few-shot machine learning
| Â | Material system | ||
|---|---|---|---|
Parameter | LSFO | STO/Ge | MoO3 |
Support classes | 2 | 3 | 3 |
Batch size | 8 | 8 | 16 |
Image size (pixels) | 2048 × 2048 | 3042 × 3044 | 512 × 512 |
Total chips | 500 | 1000 | 256 |
Chip size (pixels) | 34 × 35 | 95 × 95 | 32 × 32 |
Encoding | ResNet101 | ResNet101 | ResNet101 |
|  | pretrained = True | pretrained = True | pretrained = True |
Distance metric | Euclidean | Euclidean | Euclidean |
Similarity module | Protonet | Protonet | Protonet |
Normalization | Softmax | Softmax | Softmax |
Training | None | None | None |
Computing | device = CPU | device = CPU | device = CPU |
