Dynamic Directional Convolution Vector Field for Active Contour Models
Gang Wang, Jianming Liang, Yang Wang · 2011
In this paper, we propose a novel dynamic external force for snakes named dynamic directional convolution vector field (DDCVF). It makes use of the gradients of gray-level images and defines positive and negative boundaries in horizontal and vertical directions, respectively. Furthermore, DDCVF is calculated by convolving the user-defined vector field kernel with the edge map generated from the image in the two directions separately. Experimental results show that the DDCVF snake has a large capture range and better robustness to disturbance and initialization.