A study of software architecture for real-time image and graphic processing for time-sequenced 3-D CT images

Abstract

Software architecture is studied, in which real time image and Graphic processing are executed for time-sequenced three-dimensional computer tomogram images, to observe the organs with movement, such as heart or lung. Because such system must process so huge number of images, its observer might wait a while for processing images and then generating graphic images. In developing such system, it might be efficient to utilize GPU (graphic processor unit) for processing. The experimental result shows that the processing times on GPU are almost 1/4 of the times on CPU including memory transfer. Furthermore some suitable software architecture must be investigated in consideration of the size of targeted images and the processing time for image process and graphic process. In this paper, software architecture in the system will be proposed, which continues to display animated graphic images naturally as heart movement even while huge image process executes. © 2012 Springer-Verlag GmbH.