Cluster of microcalcifications can be an early sign of breast cancer. In this paper we present a deep convolutional neural network for microcalcification detection and compare its results to a classical approach. In this work we used 238 mammograms to train and validate our neural network to recognize which pixels in a mammogram correspond to a calcification; we tested the results on 52 images and obtained an accuracy of 83.7% against only 58% of the classical approach. Our results show how deep learning could be an effective tool to use for microcalcification detection and segmentation, outdoing classical approaches.

Keywords: Deep convolutional neural network, deep learning, segmentation, microcalcification, mammography imaging

File: https://link.springer.com/chapter/10.1007/978-981-10-5122-7_110