Mass detection on mammogram images: a first assessment of deep learning techniques

Authors

Abstract

Deep Learning approaches have gathered a lot of attention lately. In this work, we study their application to the breast cancer field, in particular for mass detection in mammograms. Several experiments were made on a real mammogram benchmark dataset. Deep Learning approaches were compared to other classification methodologies. It was concluded that, al- though useful, the implementation used does not outperform SVMs. Fur- ther study and adjustment of the method for this application is needed.

Keywords

autoencoder

Conference

19th Portuguese Conference on Pattern Recognition 2013