Title :
A new deformable contour method
Author :
Wang, Xun ; Wee, William G.
Author_Institution :
Dept. of Electr. & Comput. Eng. & Comput. Sci., Cincinnati Univ., OH, USA
Abstract :
In this paper, a new deformable contour method is presented and tested. The method incorporates several advantages and avoids the disadvantages of both the level set and the non-deterministic deformable contour methods. It takes the level set approach as a core procedure with an enhanced velocity control using a region growing-like function, and the introduction of the non-deterministic approach to generate a set of candidate contours for evaluation by an energy function. The resulting procedure can handle blurred and/or opened contours, and noisy contours in images with a complex structure. The procedure is automatic and rather insensitive to the initial location, and has only one parameter σ to adjust. Local energy minima contours can easily be tuned to attend. Experimental results on human abdominal CT images, brain MRI images, and ultrasound pig heart images are favorably reported
Keywords :
biomedical MRI; brain; cardiology; computerised tomography; edge detection; image segmentation; medical image processing; minimisation; ultrasonic imaging; abdominal CT images; blurred contours; brain MRI images; candidate contours; deformable contour method; energy function; enhanced velocity control; level set approach; local energy minima; noisy contours; non-deterministic approach; opened contours; region growing; ultrasound pig heart images; Computed tomography; Computer science; Deformable models; Heart; Image processing; Image segmentation; Level set; Magnetic resonance imaging; Stochastic processes; Testing;
Conference_Titel :
Image Analysis and Processing, 1999. Proceedings. International Conference on
Conference_Location :
Venice
Print_ISBN :
0-7695-0040-4
DOI :
10.1109/ICIAP.1999.797633