3D volume visualization system based on GPUs for medical big data

Abstract

With the appearance of computerization society, medical equipment has been closely coupling with computers. Images from medical equipment have become very useful basic data in diagnosing the state of a patent. These images have often obtained from medical computerized equipment such as X-ray system, Ultrasound system, Computerized Tomography (CT), Magnetic Resonance Imaging (MRI), and so on. For increasing the effective value of those images, we have to employ the preprocessing techniques such as a median filter, a Gaussian filter, and a mean filter. And also, we take the special algorithms such as a phase correction, a dispersion compensation, and a Hilbert transform. These additional techniques make it impossible to watch the images in a real time. Thus, we have to get a rapid processing of medical images for the real-time display of those images. In this research, we will propose a method to give a real-time display of images using multi-thread techniques and parallel processing algorithms. In a multi-thread technique, if two threads share the same resource, one thread may change the resource data during the processing of the other thread. We will design a method to avoid the collision in a resource usage right between threads.