Medical Image Display And Processing System Based On GPU Acceleration

With the rapid development of modern medicine,medical image has become an integral part of modern medicine.At the same time,medical image processing system has also become an indispensable auxiliary diagnostic tool.However,due to the diversity and complexity of the actual clinical needs,the existing open source software or commercial software can't fully meet the demand,so developing a system platform with high expansibility and flexibility is very necessary.The work of this article is divided into two parts,the development of the software platform and the research on the acceleration of GPU segmentation.According to the first people's Hospital of Yunnan province(Kunhua hospital)the actual clinical needs of tumor disease comprehensive diagnosis and treatment center,designed a software platform of ImageView,it is in the cross platform C++ application development framework,the development of a Qt has high scalability of the software development environment using Visual Studio 2015,we use VTK(Visualization Toolkit)visualization of medical images,and the use of ITK(Insight Segmentation and Registration Toolkit)to achieve the basic image processing and segmentation algorithm.Image View can present different views,including single window slice view,single sequence multi window view,interactive slice attempt and 3D rendering view.It can also display multiple views at the same time and contrast.It can also be connected to the hospital's cancer patient information collection system through the network to download the case image and process it.ImageView also has good extensibility and can expand the main functions of the software.After our test,we found that ImageView data read fast,and can read multiple 3D data at the same time.The performance of view interaction is good,and the rendering effect of 3D is realistic.In the process of our development,it is found that the traditional method of medical image segmentation has a huge amount of computation,which hinders the practical clinical application.In order to solve this problem,we use the GPU acceleration method to speed up the image segmentation algorithm.We make an in-depth analysis of the segmentation algorithm,and use the CPU+GPU cooperative processing method to speed up the algorithm.We use the OpenCL(Open Computing Language)platform to implement the parallel computing of the GPU segmentation algorithm.Experimental results show that GPU parallel computing has high speedup in processing 3D images,and can be used for fast image segmentation,which lays the foundation for practical clinical application.