Residual convolutional neural networks for breast density classification

In this paper, we propose a data-driven method to classify mammograms according to breast density in BIRADS standard. About 2000 mammographic exams have been collected from the "Azienda Ospedaliero- Universitaria Pisana" (AOUP, Pisa, IT). The dataset has been classified according to breast density in the BI-RADS standard. Once the dataset has been labeled by a radiologist, we proceeded by building a Residual Neural Network in order to classify breast density in two ways. First, we classified mammograms using two "super-classes" that are dense and non-dense breast. Second, we trained the residual neural network to classify mammograms according to the four classes of the BI-RADS standard. We evaluated the performance in terms of the accuracy and we obtained very good results compared to other works on similar classification tasks. In the near future, we are going to improve the results by increasing the computing power, by improving the quality of the ground truth and by increasing the number of samples in the dataset.