Medical Image Classification on SIIM-COVID (test)
52.66AccuracyEG-AL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EG-ALSampling Strategy=Explainability-Guided, Total Labeled Samples=570, Active Learning Rounds=7, Backbone=DenseNet-121, Pre-training=ImageNet, Evaluation Protocol=5-shot episodes, Number of Seeds=52026.02 | 52.66 | 66.92 | |
| RandomSampling Strategy=Random, Total Labeled Samples=570, Active Learning Rounds=7, Backbone=DenseNet-121, Pre-training=ImageNet, Evaluation Protocol=5-shot episodes, Number of Seeds=52026.02 | 38.28 | 54.21 | |
| BaselineSampling Strategy=Random (initialization), Total Labeled Samples=150, Active Learning Rounds=0, Backbone=DenseNet-121, Pre-training=ImageNet, Evaluation Protocol=5-shot episodes, Number of Seeds=52026.02 | 30.32 | 56.74 |