Table 12 Results on the validation data set considering different data augmentation methods
From: Automated classification of polyps using deep learning architectures and few-shot learning
Model | Data augmentation | Acc | ||
|---|---|---|---|---|
| Â | random flip | random rotation | random contrast | Â |
BiT-R152x4 | Â | Â | Â | 0.8155 |
\(\checkmark\) | Â | Â | 0.8213 | |
\(\checkmark\) | Â | \(\checkmark\) | 0.4543 | |
\(\checkmark\) | \(\checkmark\) | Â | 0.7968 | |
\(\checkmark\) | \(\checkmark\) | \(\checkmark\) | 0.4469 | |
EfficientNet-B7 | Â | Â | Â | 0.7551 |
\(\checkmark\) | Â | Â | 0.7903 | |
\(\checkmark\) | Â | \(\checkmark\) | 0.7936 | |
\(\checkmark\) | \(\checkmark\) | Â | 0.8091 | |
\(\checkmark\) | \(\checkmark\) | \(\checkmark\) | 0.8212 | |
Ours | Â | Â | Â | 0.7930 |
\(\checkmark\) | Â | Â | 0.8950 | |
\(\checkmark\) | Â | \(\checkmark\) | 0.8210 | |
\(\checkmark\) | \(\checkmark\) | Â | 0.8242 | |
\(\checkmark\) | \(\checkmark\) | \(\checkmark\) | 0.6016 | |