Mitosis Detection using Convolutional Neural Network 

IEEE Research Paper : 
http://ieeexplore.ieee.org/document/7846429/

Dataset :
MITOS dataset provided in ICPR 2014
https://mitos-atypia-14.grand-challenge.org/donwload/

Summary : 

	From the dataset each slide image is examined to test for mitotic and non-mitotic cell. In the preprocess step, each cell is extracted from the slide ignoring the non-cellular structures. Background subtraction, K-Means Clustering and Blob Analysis concepts are implemented to extract the cellular structures. After the preprocess step the extracted cell patches are sent to a convolutional neural network to extract the desired features. The test images are classified based on the extracted features of CNN.
	CNN has been implemented using Keras library - Theano backend