Histopathologic-Cancer-Detection

Detection of cancer in histopathologic slides of the mammary tissue using Computer Vision The main aim of this project is to find patches of cancerous cells in histopathologic slides of the mammary tissue. We performed several experiments on the data by changing the hyperparameters such as Activation function, Number of Hidden layers, Optimiser, Loss function and Learning Rate. We achieved accuracy of 94.6% by using Convolutional Neural Networks with ADAM optimiser, ReLU activation, Binary Cross Entropy and a learning rate of 0.0001 The data can be found at : https://www.kaggle.com/c/histopathologic-cancer-detection