DocumentCode :
1807680
Title :
Automated detection of left ventricular epi- and endocardial contours in short-axis MR images
Author :
van der Geest, Rob J. ; Jansen, E. ; Buller, VGM ; Reiber, JHC
Author_Institution :
Lab. for Clinical & Exp. Image Process., Univ. Hospital, Leiden, Netherlands
fYear :
1994
fDate :
25-28 Sep 1994
Firstpage :
33
Lastpage :
36
Abstract :
Quantitative analysis of global and regional left ventricular function from short-axis multi-slice Cine MR imaging studies currently requires manual tracing of endo- and epicardial contours in many images. To circumvent the limitations which are associated with manual contour tracing procedures, a software package (MASS: MR Analytical Software System) has been developed with automated contour detection. The contour detection follows a model-based approach incorporating the assumption that contours and images exhibit a gradual change from phase to phase and from slice to slice. The user may interact with the automated contour detection to control the analysis procedure and to avoid the potential risk of error propagation. Compared to manually obtained results, the contour detection software accurately assessed endocardial volumes, epicardial volumes and ejection fraction (correlation coefficients: 0.90-0.97). Total average analysis time for a complete imaging study was 25 minutes
Keywords :
biomedical NMR; cardiology; edge detection; medical image processing; muscle; MASS; MR Analytical Software System; automated contour detection; correlation coefficients; ejection fraction; endocardial contours; endocardial volumes; epicardial volumes; error propagation; global left ventricular function; left ventricular epicardial contours; model-based approach; regional left ventricular function; short-axis MR images; short-axis multi-slice Cine MR imaging; software package; Blood; Detection algorithms; Image analysis; Lungs; Myocardium; Phase detection; Risk analysis; Software packages; Software systems; Weight control;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computers in Cardiology 1994
Conference_Location :
Bethesda, MD
Print_ISBN :
0-8186-6570-X
Type :
conf
DOI :
10.1109/CIC.1994.470256
Filename :
470256
Link To Document :
بازگشت