Abstract
This paper describes a development-supporting environment of general-purpose computing on graphics processing units (GPGPU) programming for Insight Segmentation and Registration toolkit (ITK). Performance of CT scanners has improved in recent years, thus amount of information in a CT image is becoming large. To solve the problem of the increasing computational cost for computer-aided diagnosis (CAD), the acceleration of the computation is indispensable. The authors have proposed a method to accelerate arbitrary image processing programs by GPU parallel computation for ITK. In this paper, we present the mechanism of proposed method, usage examples and discussion about the effect of acceleration.
