-Mole_classification

DNN, which classifies moles into three classes, using the resnet and the transfer learning method. 3 classes of moles represented by nevus, keratosis, melanoma

Important libraries:

Tensorflow, keras, scikit-learn(sklearn)

Data:

2000 images train dataset

300 images valid dataset

600 images test dataset

Epoch:

20

Result:

~69% accuracy on train

~65.5% accuracy on test