A semi-supervised fuzzy GrowCut algorithm to segment and classify regions of interest of mammographic images
作者：
Highlights：
• We propose a Fuzzy semi-supervised version of the GrowCut algorithm.
• We reduced dependence of GrowCut on user experience, using simulated annealing.
• To improve robustness to point selection, we modified the GrowCut evolution rule.
• We evaluated our approach by classifying 685 digital mammograms.
• Our approach could reach an overall accuracy of 91% for fat tissues.
摘要
•We propose a Fuzzy semi-supervised version of the GrowCut algorithm.•We reduced dependence of GrowCut on user experience, using simulated annealing.•To improve robustness to point selection, we modified the GrowCut evolution rule.•We evaluated our approach by classifying 685 digital mammograms.•Our approach could reach an overall accuracy of 91% for fat tissues.
论文关键词：Breast cancer,Mammographic image analysis,Semi-supervised image segmentation,GrowCut algorithm,Fuzzy segmentation,Simulated annealing
论文评审过程：Received 1 March 2015, Revised 13 May 2016, Accepted 2 August 2016, Available online 3 August 2016, Version of Record 16 August 2016.
论文官网地址：https://doi.org/10.1016/j.eswa.2016.08.016
