3D automated colon segmentation for efficient polyp detection

Author

Ismail, Mahamod ; Elhabian, Shireen ; Farag, Aly ; Dryden, G. ; Seow, A.

Author_Institution

Univ. of Louisville, Louisville, KY, USA

fYear

2012

fDate

20-22 Dec. 2012

Firstpage

48

Lastpage

51

Abstract

With polyps being the main cause of colorectal cancer, accurate colon segmentation is a crucial step for polyp detection in a virtual colonoscopy system. This paper presents a fully automated segmentation framework for the colon which is based on convex formulation of the active contour model. Our approach is tested on 7 sets where the results are further validated for polyp detection. Results show the efficiency of the framework with an overall accuracy of 99%, and high sensitivity of polyp detection.

Keywords

biomedical optical imaging; cancer; endoscopes; image segmentation; medical image processing; tumours; 3D automated colon segmentation; active contour model; colorectal cancer; convex formulation; fully automated segmentation framework; polyp detection; sensitivity; virtual colonoscopy system; Accuracy; Active contours; Cancer; Colon; Indexes; Shape; Silicon; Mumford-Shah; convexification; haustral folds; polyps; region growing; shape index;

fLanguage

English

Publisher

ieee

Conference_Titel

Biomedical Engineering Conference (CIBEC), 2012 Cairo International

Conference_Location

Giza

ISSN

2156-6097

Print_ISBN

978-1-4673-2800-5

Type

conf

DOI

10.1109/CIBEC.2012.6473334

Filename

6473334