VOLES: Vascularity-Oriented LEvel Set algorithm for pulmonary vessel segmentation in image guided intervention therapy

Delicate surgical planning and accurate guidance plays an important role in successful image guided intervention. In interventional lung cancer diagnosis and treatments, precise segmentation of pulmonary vessels from lung CT images provides vital visualization for pre-op planning and inra-op guidance to avoid major vessel damage. While simple thresholding and window/level setting can briefly segment different tissues, their results are not accurate. Recently, level set methods have been increasingly and successfully used in various organ segmentations, however, the penalty on large curvature makes the evolution along vascular structure slow, thus rendering difficulty in lung vessels. In this paper, we propose a Vascularity-Oriented LEvel Set algorithm (VOLES) to offset the curvature effect on the evolving front along vessel directions, also the evolution direction can be adaptively adjusted based on the joint intensity and vesselness statistics to prevent leakage and to adapt to intensity inhomogeneity. The VOLES algorithm is validated using lung CT images in the experiments, and results show it outperforms the traditional level set method on pulmonary vessel segmentation.