Convolutional Autoencoder Application for Breast Cancer Classification

There are many related works for breast cancer detection using convolutional neural networks (CNN), and most of them rely only on feature extraction after the convolutions. However, because of the huge number of parameters in the models, the time of computation will be increased. In the current work, a convolutional autoencoder was proposed to reduce the complexity of the model, the number of parameters and as a result, prevent overfitting the model. In this p aper, there were trained several autoencoder models from scratch with different architectures and different hyperparameters. Moreover, 37% noise was applied on the input images and the model could reconstruct the original image with 84.72% accuracy. The sensitivity and precision of the proposed model were 86.87% and 80.23% respectively.