Synthetically Generated Data : https://drive.google.com/file/d/1AgL6xkahmx19BFhQ0r8ESmfVuMyvWOFA/view?usp=sharing
InBreast Dataset : https://drive.google.com/file/d/1knX6qcx3tEXauYQAdV_Da3ew0VxSa4_g/view?usp=sharing
The test set contains augmented images. Remember to remove them before using
To generate images please go to the Pro-GAN folder and run the train.py file. You can modify the hyper-parameters by changing the config.py file.
Run all the cells in autoencoder-reconstruction notebook. It will train the model and save the best performing models weights.
Run all the cells in autoencoder-classification notebook. It will train the model based on the proposed approach.
Additionally, the autoencoder-classification model can be found in model.py file. main_DNN.ipynb contains codes to run the off-the-self DNN models for comparison.