Ankle cartilage surface segmentation using directional gradient vector flow snakes

Abstract

Ankle cartilage surface segmentation is a prestep to compute the cartilage thickness which is used to evaluate the treatment of ankle joint diseases. Because of the un-ideal imaging condition, the acquired ankle images are often noised and the cartilage parts in the images are often disconnected, thus traditional techniques such as gradient-based edge detection are not suitable for the segmentation of these surfaces. In this paper, we developed an active contour model based ankle cartilage surface segmentation algorithm. The algorithm used gradient vector flow (GVF) as external forces. In order to make the GVF snake more stable and to converge at the correct surfaces, directional gradient is used to produce the gradient vector flow field. Experimental results indicate that the new approach utilizing directional gradient is more robust than the traditional GVF snake.