This respository contains 2 notebooks with code to tackle the (closed) kaggle competition Histopathological cancer detection. It consists on a binary classification folder with a balanced data set. It is my first approach to hands-on deep learning. I have used Convolutional Neural Networks with Keras as deep learning programming framework.
The accuracy is far from good (also limited by resources: no GPU available), but you can take it as a starting point. I think it is clear and easy to follow for deep-learning beginners.
#IMPORTANT:
Images in training and test set must be located in subdirectories: one subdirectory ("images") in test directory and as many subdirectories as classes in training directory.
The training images matching a specific class must be located in the corresponding folder.
In this project there are 2 classes (binary classification problem) and the structure in data directory was the following:
