Proc. SPIE. 6914, Medical Imaging 2008: Image Processing
KEYWORDS: Human-machine interfaces, 3D acquisition, Optical spheres, Computer aided diagnosis and therapy, Image segmentation, Pancreas, Medical imaging, Computed tomography, 3D displays, 3D image processing
Automatic segmentation of specific organ from CT image is one of the most important problems for computer aided diagnosis. Therefore, many automatic segmentation methods have been proposed. However, in these conventional methods, it is necessary to set some empirical parameters by users for supporting the automatic segmentation. It takes, however, much time to obtain the optimum parameters for segmentation. Another problem is that although the optimum value of parameters varies spatially, conventional segmentation methods can give only one value of parameters for whole organ. In this paper, we propose a new segmentation method named hybrid segmentation of CT image based on level set segmentation. A local control parameter which can be controlled by the interactive interface used PHANToM omni is introduced for segmentation of pancreas from three dimensional CT image. It is shown that the proposed method can increase accuracy in less time than hand-written segmentation.