A hybrid active contour model driven by novel global and local fitting energies for image segmentation

摘要

This paper proposes a hybrid active contour model driven by novel global and local fitting energies for image segmentation. First, the global fitting term is defined by minimizing the difference between the weighted global fitted image and the original image, which describes the global information more accurately. Second, the local fitting term is defined by the cross entropy to compute the local information, which improves the generality of the proposed model. Experiments are performed on synthetic and real images and the results demonstrate that compared with existing active contour models, the proposed model has advantages in terms of segmentation performance and robustness to initial contour and noise.