DocumentCode :
3299573
Title :
A video-frame based registration using segmentation and graph connectivity for Wireless Capsule Endoscopy
Author :
Karargyris, Alexandros ; Bourbakis, Nikolaos
Author_Institution :
Coll. of Eng., Wright State Univ., Dayton, OH
fYear :
2009
fDate :
9-10 April 2009
Firstpage :
74
Lastpage :
79
Abstract :
Gastroenterologists using wireless capsule endoscopy (WCE) extract valuable information regarding the health condition of the intestine. WCE produces, however, near 50,000 frames that have to be examined. In this paper a registration methodology is presented to reduce the amount of frames, which are the same or similar. It utilizes a segmentation scheme and a graph method to register similar regions in two consecutive frames. In addition our methodology offers a degree of similarity between two frames using a confidence function of similarity, in order to discard extremely dissimilar frames and thus set a mark of a new series of frames. Finally, a useful feature called Center of Image Similarity is also presented. Illustrative results from the methodology are also given in this paper.
Keywords :
endoscopes; medical image processing; gastroenterologists; graph connectivity; image similarity; intestine; segmentation; video registration; wireless capsule endoscopy; Cameras; Computer vision; Data mining; Digestive system; Educational institutions; Endoscopes; Image registration; Intestines; Layout; Shape; Confidence function; Connectivity Region Graphs; Registration; Spatial Similarity; Wireless Capsule Endoscopic Imaging;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Life Science Systems and Applications Workshop, 2009. LiSSA 2009. IEEE/NIH
Conference_Location :
Bethesda, MD
Print_ISBN :
978-1-4244-4292-8
Electronic_ISBN :
978-1-4244-4293-5
Type :
conf
DOI :
10.1109/LISSA.2009.4906713
Filename :
4906713
Link To Document :
بازگشت