Improved gradient vector flow snake algorithm for lateral ventricle segmentation
Jingli Zhang · Zhongguo yixue wulixue zazhi · 2015
To improve the accuracy and robustness of the lateral ventricle segmentation by improving gradient vector flow snake(GVF Snake) algorithm. The prior information was firstly applied by the algorithm to initialize and automate contours.And then, the brain images were processed by edge detection operator to get the edge maps. Finally, GVF fields were calculated based on the edge maps, and the lateral ventricle segmentation results were obtained by contour evolution. Different lateral ventricle images were selected to experimentally verify the algorithm. The improved algorithm can accurately extract the contour of lateral ventricle sharp angular points, overcoming the over- convergence of GVF Snake. And the improved algorithm can achieve better segmentation results when the junction between left and right lateral ventricle is small.