Image Structure Adaptive Gradient Vector Flow for Active Contours

The gradient vector flow (GVF) is an effective external force for active contours, but it´s smoothness constraint doesn´t take any image structure into account and the resulting diffusion equation is isotropic which would oversmooth weak edges. In this article, a novel external force named image structure adaptive gradient vector flow (ISAGVF) is proposed for active contours. The proposed ISAGVF employs certain smoothness constraint which is able to be adaptive to image structure, such that the ISAGVF snake would outperform the GVF snake in terms of noise robustness, weak edge preserving and narrow and deep concavity convergence, while possessing some other desirable properties such as enlarged capture range and insensitivity to initialization. Experiments on synthetic and real images manifest these properties of ISAGVF snake.