Table 3 Influence of different label scales on model effects
From: A dynamic few-shot learning framework for medical image stream mining based on self-training
Dataset | Label proportion (%) | Mean |
|---|---|---|
Abdomen-MRI | 5 | 72.26 |
10 | 72.53 | |
20 | 72.61 | |
50 | 74.28 | |
Abdomen-CT | 5 | 61.55 |
10 | 62.06 | |
20 | 62.93 | |
50 | 63.42 |