An Interactive Segmentation Based on CT Image Series

摘要

Segmentation is an essential and important step in the analysis of medical images. This paper proposes an interactive scheme for segmentation of 3D CT image series data. User starts with an improved live wire algorithm on individually selected slices. If segmented contours on at least two slices are available, all contours on slices in between can be computed utilizing contour interpolation algorithm. If the variations of the object boundary between user-selected slices are significant, automatic optimization algorithm is applied for reduction of the variations in the interpolated contours. Experiments show that our scheme can obtain the region boundary of interest on a series of CT images quickly and reliably with only little user intervention.