Convolutional Neural Networks for mammogram classification in BIRADS standard: development and preliminary tests

Francesca Lizzi, Paolo Bosco, Davide Caramella, Carolina Marini, Alessandra Retico, Antonio Traino, Maria Evelina Fantacci

Apr 11, 2018 (modified: May 16, 2018) MIDL 2018 Abstract Submission readers: everyone
  • Abstract: The aim of our project is to transpose the European Directive 59/2013/EURATOM which regulates the use of ionizing radiation in medicine and the information provided to patients on radiation dose. Controlling the dose is a really important aspect when we deal with breast cancer screening because breast is made of a radio-sensitive tissue and screening programs expose to radiation a lot of healthy women. In order to personalize the dose index, it is necessary to take into account breast density because to have a sufficient sensitivity a higher dose is given to the patient. For this reason we propose to classify mammograms with a Convolutional Neural Network (CNN) in BIRADS standard. To assess density in screening program, an automatic algorithm is necessary because of the large number of women who partecipates.
  • Keywords: Convolutional Neural Networks, Breast Density, BIRADS, Dose Control
  • Author affiliation: National Institute for Nuclear Physics (INFN), University of Pisa, Azienda Ospedaliero-Universitaria Pisana (AOUP)
0 Replies

Loading