Unifying Boundary and Region-Based Information for Fuzzy-Based Active Region Tracking

In this paper we present a novel unifying boundary and region-based information under a fuzzy-based active region framework for tracking. The boundary information is determined by the geodesic model which finds the best minimal-length smooth curve for a fuzzy map to the difference frame, and the region information indicating the motion-based region properties is estimated by a class measure based on the fuzzy c-mean clustering. Here, we define a novel class measure and boundary membership. The defined objective function is minimized using a gradient descent method where a level sets approach is used to implement the resulting PDE system. And satisfactory results of this method are obtained in real sequences.