CARTILAGE SURFACE TRACKING USING DIRECTIONAL GRADIENT VECTOR FLOW SNAKES (WA-P4)

Author(s) : Jinshan Tang (Department of Electrical Engineering,University of Virginia, USA) Steven Millington (Center for Applied Biomechanics, #Department of Orthopaedic Surgery,University of Virginia, USA) Scott T Acton (Department of Electrical Engineering,University of Virginia, USA)

Abstract : The accuracy of the surface extraction and segmentation of magnetic resonance images of highly congruent joints with thin articular cartilage layers have a significant effect on the percentage errors and reproducibility of quantitative measurements (e.g. thickness and volume) of the articular cartilage. Traditional techniques such as gradient-based edge detection are not suitable for the segmentation of these surfaces. This paper describes the segmentation of articular cartilage surfaces using an active contour model with a gradient vector flow (GVF) based external force. In order to make the GVF snake more stable and to converge at the correct surfaces, gradient direction information is used to produce the gradient vector flow field. Experimental results indicate that the new approach utilizing the gradient direction information is more robust than the traditional GVF snake.

Menu