| |
| In this paper we present a novel Uncoupled Fuzzy-based Geodesic Active Regions (UFGAR) framework for dealing with frame partition problems in image segmentation. Like the usually Geodesic Active Regions (GAR) framework presented by N. Paragios and R. Derichhave, our framework unifies boundary and region-based information. The boundary information of the mixture model is determined using a edge detector based on the fuzzy membership functions, and the region information indicates the region intensity properties is estimated by a measure based on the fuzzy membership functions too. The defined objective function is minimized using a gradient descent method where a level set approach is used to implement the resulting PDE system. But unlike GAR, which is based on the Maximum Likelihood Principle for the observed density function image histogram using a mixture of Gaussian elements, our model is based on FCM and fuzzy membership functions. We define a novel region measure based on fuzzy membership functions to measure the property of regions. According to the motion equation, the initial curve is propagated toward the segmentation result under the influence of boundary and region-based segmentation forces, and being constrained by a regularity force. The performance of this method is demonstrated by a synthetic image, and the results of its application to brain images are presented. |
|
| Keywords:Image segmentation,Geodesic Active Regions,FCM, fuzzy membership function,level set |
|
