Intravascular Ultrasound image segmentation: A helical active contour method

During an Intravascular Ultrasound (IVUS) examination, a catheter with an ultrasound transducer is introduced in the body through a blood vessel and then pulled back to image a sequence of vessel cross-sections. An IVUS exam results in several hundred noisy images often hard to analyze. Hence, developing powerful automatic analysis tool would facilitate the interpretation of structures in IVUS images. In this paper we present a new IVUS segmentation method based on an original active contour model. The contour has a helical geometry and evolves like a spiral shape that is distorted until it reaches the artery lumen boundaries. Despite the use of a simple statistical model and a very sparse initialization of the snake, the algorithm converges to satisfying solutions that can be compared with much more sophisticated segmentation methods. To validate the method, we compared our results to manually traced contours and obtained an Hausdorff distance < 0∶61mm (n = 540 images) indicating the robustness of the method.