Research And Application On Algorithms Of The Medical Images Based On GPU

Three-dimensional reconstruction and processing of medical image is importanceapplication of computerized technology, it is combination of computer visiualizationtechnology and image processing technology. Calculating works of three-dimensionalreconstruction about medical images is very large, and the requirement about GPU(Graphics Processing Unit) and other hardware resources is high, which leads to highequipment costs.The broad applications of GPU make it possible to reduce cost, andGPU is more efficient by far than CPU in the processing of graphic image data andcomplex algorithms.The thesis analyzes volume rendering algorithm, According to the features ofmedical image and GPU, Optimizing the raycasting on GPU to render the medicalimage. Implemented based GPU ray casting algorithm and part of the image processingfunctions. According to the characteristics of GPU, selected CUDA implementationplatform, design and implementation is divided into GPU and CPU client side modules.Module in the GPU side, to realize an improved algorithm for ray casting. the originalmedical image is processed, the effect is greatly improved, higher image quality can beachieved by real-time, and the image filtering, sharpening and the other processing. Afertesting, the reaults of this project meet the design requirements.