Options
2015
Conference Paper
Titel
Sliding Level Set-Based Boundary: Fully Automated Dense Breast Segmentation in Native MR Mammograms
Abstract
In this work, we present a fully automated method for accurate breast tissue segmentation, which is the fundamental step for automated breast density evaluation in MR data. The approach is designed to tackle challenging dense breast cases, where the parenchymal tissue is adjacent to the pectoral muscle tissue and the clear breast-muscle boundary is missing. The approach consists of three steps: simultaneous intensity inhomogeneity correction and segmentation, extraction of the breast and restoration of the breast-muscle boundary. The method was tested on 20 breast datasets and compared to the state-of-the-art technique. We achieved an overall average Dice Similarity Coefficient (DSC) of 96.7 ± 1.14%.