Adversarial deep structured nets for mass segmentation from mammograms
TL
TL;DR
A novel end-to-end network for mammographic mass segmentation is proposed which employs a fully convolutional network (FCN) to model a potential function, followed by a conditional random field (CRF) to perform structured learning.
Abstract
Mass segmentation provides effective morphological features which are important for mass diagnosis. In this work, we propose a novel end-to-end network for mammographic mass segmentation which employs a fully convolutional network (FCN) to model a potential function, followed by a conditional random field (CRF) to perform structured learning. Because the mass distribution varies greatly with pixel position, the FCN is combined with a position priori. Further, we employ adversarial training to eliminate over-fitting due to the small sizes of mammogram datasets. Multi-scale FCN is employed to improve the segmentation performance. Experimental results on two public datasets, IN breast and DDSM-BCRP, demonstrate that our end-to-end network achieves better performance than state-of-the-art approaches.1
References19 items
1
Automatic Liver Segmentation Using an Adversarial Image-to-Image Network
2
Deep Multi-instance Networks with Sparse Label Assignment for Whole Mammogram Classification
3
Mammographic Mass Segmentation with Online Learned Shape and Appearance Priors
4
Guest Editorial Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique
5