Adversarial deep structured nets for mass segmentation from mammograms
Authors:
- 1 April 2018
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 847-850
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.Keywords
All Related Versions
This publication has 14 references indexed in Scilit:
- Deep Multi-instance Networks with Sparse Label Assignment for Whole Mammogram Classification2017
- Mammographic Mass Segmentation with Online Learned Shape and Appearance Priors2016
- Conditional Random Fields as Recurrent Neural Networks2015
- Deep Learning and Structured Prediction for the Segmentation of Mass in Mammograms2015
- Deep structured learning for mass segmentation from mammograms2015
- Tree RE-weighted belief propagation using deep learning potentials for mass segmentation from mammograms2015
- INbreastAcademic Radiology, 2012
- Robust Higher Order Potentials for Enforcing Label ConsistencyInternational Journal of Computer Vision, 2009
- Digital Mammographic Computer Aided Diagnosis (CAD) using Adaptive Level Set Segmentation2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2007
- Current Status of the Digital Database for Screening Mammography1998