计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (20): 166-172.DOI: 10.3778/j.issn.1002-8331.1604-0398

• 图形图像处理 • 上一篇    下一篇

基于多核并行化差异进化算法的图像配准方法

曹国刚,张  晴,张培君,王志敏   

  1. 上海应用技术大学 计算机科学与信息工程学院,上海 201418
  • 出版日期:2017-10-15 发布日期:2017-10-31

Multicore-based parallelized differential evolution for image registration

CAO Guogang, ZHANG Qing, ZHANG Peijun, WANG Zhimin   

  1. School of Computer Science & Information Engineering, Shanghai Institute of Technology, Shanghai 201418, China
  • Online:2017-10-15 Published:2017-10-31

摘要: 随着多核处理器的逐渐普及,它将成为未来处理器发展的必然趋势,各种应用都将采用多核架构。在基于差异进化算法的基础上,在图形工作站上充分利用多核处理器的并行计算能力,实现多核并行化差异进化算法,并将其应用于三维医学图像配准,实现快速配准算法。实验数据表明,与传统方法相比,多核并行化方法在保证精度和稳定性的前提下,配准速度提升接近于理想加速比。

关键词: 差异进化, 多核处理器, 并行化, 图像配准

Abstract: With the popularization of multicore CPU, it will be the trend of CPU and be used to different applications. A multicore-based parallelized differential evolution algorithm is proposed, which takes advantage of parallel computing capabilities of multicore CPU on graphic workstation. It accomplishes the parallel evolution algorithm and realizes the multicore-based parallelized fast medical image registration. The experimental results show that, it not only has no difference in precision and stability with traditional image registration based on differential evolution, but also exhibits fast convergence and almost has an ideal speedup.

Key words: differential evolution, multicore Central Processing Unit(CPU), parallelization, image registration