Hybrid Active Contour Model for Segmentation of Synthetic and Real Images

Level set models are extensively used for image segmentation because of their capability to handle topological changes. In this paper, the proposed model uses combined local image information and global image information to evolve the con-tour around the object boundary, making it robust, irrespective of the inhomogeneity. The proposed model is capable to deal with bias conditions, such as intensity inhomogeneity and light effects. We test this model on synthetic, and real images, confirming its superiority over previous models.