- Tensorflow
- Keras
original dataset -http://brainiac2.mit.edu/isbi_challenge/
The data for training contains 30 512*512 images,
Data augmentation : ImageDataGenerator in keras.preprocessing.image
Output from the network is a 512*512 which represents mask that should be learned.
Sigmoid activation function makes sure that mask pixels are in [0, 1] range.
Epochs = 5
Accuracy ~ 0.95.
Loss function - binary crossentropy.