Title :
A hybrid approach for hemorrhage segmentation in pelvic CT scans
Author :
Davuluri, Pavani ; Tang, Yang ; Wu, Jie ; Ward, Kevin R. ; Belle, Ashwin ; Najarian, Kayvan ; Cockrell, Charles H. ; Hargraves, Rosalyn H.
Author_Institution :
Dept. of Electr. & Comput. Eng., Virginia Commonwealth Univ., Richmond, VA, USA
Abstract :
Hemorrhage is the leading cause of death in patients with severe pelvic fractures within the first 24 hours after the injury. Hence, it is vital for physicians to quickly identify hemorrhage and assess bleeding severity. However, it is rather time consuming for physicians to evaluate all the CT images. Therefore, an automated hemorrhage segmentation system is needed to assist physicians. This paper proposes a hybrid approach for hemorrhage segmentation from pelvic CT scans. This approach utilizes region growing technique with integration of contrast information from the previous and subsequent slices. The results show that the method is able to segment hemorrhage well with acceptable results. Hemorrhage volume is also determined. A statistical t-test is conducted to determine if the calculated hemorrhage volume using the proposed method is significantly different from the manually detected volume.
Keywords :
computerised tomography; image segmentation; injuries; medical image processing; bleeding severity assesment; contrast information; death; hemorrhage segmentation; injury; pelvic CT scans; region growing technique; severe pelvic fractures; Arteries; Bones; Computed tomography; Educational institutions; Electronic mail; Hemorrhaging; Image segmentation; hemorrhage; pelvic CT images; region growing; segmentation;
Conference_Titel :
Bioinformatics and Biomedicine Workshops (BIBMW), 2011 IEEE International Conference on
Conference_Location :
Atlanta, GA
Print_ISBN :
978-1-4577-1612-6
DOI :
10.1109/BIBMW.2011.6112428