Segmentation of Tubular Structures Using Flux Maximizing Flows: An Application to Color Medical Images

摘要

The segmentation of vasculature is of utmost importance in medical image processing. It is arduous to represent the tubular structures of images obtained by recent imaging modalities with direct visualization. In this paper, we present a method of vessel segmentation in color volumetric medical images. Differ from using intensity images as data set, the improvement of traditional flux maximizing flows produce effective segmentations in color images. The visualization of vessel segmentation can illuminate the spatial structures of vasculature, making our techniques feasible for clinical diagnosis and many other applications.