Transfer learning with ResNet50 as based model.
The model classify edoscopic images into 2 grades of gastroesophageal reflux disease: Grade A and Grade B
Private dataset from Hanoi Medical University Hospital. For privacy statement, I cannot public this dataset
Because of data scarcity, I used both offline and online data augmentation to create more data
precision recall f1-score support
0 0.86 0.76 0.81 25
1 0.79 0.88 0.83 25
accuracy 0.82 50
macro avg 0.82 0.82 0.82 50
weighted avg 0.82 0.82 0.82 50