| Popis: |
Image segmentation is a complex mathematical problem, especially for images that contain intensity inhomogeneity and tightly packed objects. Magnetic Resonance (MR) muscle images often contain both of these issues, making muscle segmentation especially difficult. In this paper we propose a semi-automatic approach to segment the three muscle groups of the human thigh from MR images. The approach uses a geometric flow that incorporates a reproducing kernel Hilbert space (RKHS) edge descriptor and a geodesic distance penalty term from a set of markers and anti-markers. A method for automating the placement of the markers using atlas-based segmentation is also explored. To help deal with the intensity inhomogeneity, a new approach to estimate the bias field using a fat fraction image, called Prior Bias-Corrected Fuzzy C-means (PBCFCM), is introduced. The proposed approach was tested on five subjects, resulting in Dice coefficients above 90%. |
