上海交通大学学报(自然版)

• 自动化技术、计算机技术 • 上一篇    下一篇

基于图形处理器硬件加速的高精度医学图像融合算法

徐显,杨杰   

  1. (上海交通大学 图像处理与模式识别研究所, 上海 200240)
  • 收稿日期:2009-04-14 修回日期:1900-01-01 出版日期:2010-02-26 发布日期:2010-02-26

High Degree of Accuracy Medical Image Fusion Algorithm Based on GPU Hardware Acceleration

XU Xian,YANG Jie   

  1. (Institute of Image Processing and Pattern Recognition, Shanghai Jiaotong University, Shanghai 200240, China)
  • Received:2009-04-14 Revised:1900-01-01 Online:2010-02-26 Published:2010-02-26

摘要: 提出一种基于图形处理器(GPU)硬件加速的频域非下采样轮廓波变换(FNSCT)算法.该算法构造了更加简单、快速的频域非下采样轮廓波变换,有效消除了传统小波变换以及轮廓波变换应用于图像融合算法时引起的振铃和伪吉布斯现象.结合GPU在并行大规模浮点数及快速傅里叶变换(FFT)上的高速运算能力,解决了非下采样轮廓波变换(NSCT)速度慢的问题,实现了一种高精度的医学图像融合加速算法.

关键词: 非下采样轮廓波变换, 图形处理器, 非下采样滤波, 频域滤波, 医学图像融合

Abstract: This paper proposed a frequency domain nonsubsampled contourlet transform (FNSCT) algorithm accelerated by graphic processing unit (GPU). This approach builds a simple and highspeed nonsubsampled contourlet transform in frequency domain to eliminate the ringing and pseudoGibbs phenomena caused by wavelet or contourlet transformation. Also it accelerates the time domain nonsubsampled contourlet transform (NSCT) algorithm with the help of GPU to achieve a high speed and accuracy medical image fusion method.

中图分类号: