Premium
Centerline‐based colon segmentation for CT colonography
Author(s) -
Frimmel Hans,
Näppi J.,
Yoshida H.
Publication year - 2005
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.1990288
Subject(s) - medical imaging , medicine , computer science , segmentation , computed tomography , radiology , nuclear medicine , medical physics , artificial intelligence , computer vision
We have developed a fully automated algorithm for colon segmentation, centerline‐based segmentation (CBS), which is faster than any of the previously presented segmentation algorithms, but also has high sensitivity as well as high specificity. The algorithm first thresholds a set of unprocessed CT slices. Outer air is removed, after which a bounding box is computed. A centerline is computed for all remaining regions in the thresholded volume, disregarding segments related to extracolonic structures. Centerline segments are connected, after which the anatomy‐based removal of segments representing extracolonic structures occurs. Segments related to the remaining centerline are locally region grown, and the colonic wall is found by dilation. Shape‐based interpolation provides an isotropic mask. For 38 CT datasets, CBS was compared with the knowledge‐guided segmentation (KGS) algorithm for sensitivity and specificity. With use of a 1.5 GHz AMD Athlon‐based PC, the average computation time for the segmentation was 14.8 s. The sensitivity was, on average, 96%, and the specificity was 99%. A total of 21% of the voxels segmented by KGS, of which 96% represented extracolonic structures and 4% represented the colon, were removed.