Author Login Editor-in-Chief Peer Review Editor Work Office Work

Computer Engineering

Previous Articles     Next Articles

Real-time Ultrasound Elastic Imaging Algorithm Based on GPU Parallel Particle Swarm Optimization

YANG Xianfeng,LI Yingjie,LAI Junliang,PENG Bo   

  1. (School of Computer Science,Southwest Petroleum University,Chengdu 610500,China)
  • Received:2014-10-28 Online:2015-12-15 Published:2015-12-15

基于GPU并行粒子群优化的超声弹性实时成像算法

杨先凤,李映洁,赖俊良,彭博   

  1. (西南石油大学计算机科学学院,成都 610500)
  • 作者简介:杨先风(1974-),女,教授、硕士,主研方向:数字图像处理,数据库技术;李映洁、赖俊良,硕士;彭博(通讯作者),博士。
  • 基金资助:
    四川省教育厅基金资助重点项目(12ZA195);西南石油大学科研启航基金资助项目(2014QHZ023)。

Abstract: In order to improve the quality of elastic imaging and satisfy the real-time imaging at the same time,a Graphic Processing Unit(GPU) paralleling Particle Swarm Optimization(PSO) of real-time ultrasound elastic imaging is investigated.It describes the PSO estimating displacement of each estimated point,and uses the GPU parallel implement the method.Experimental results illustrate that the method based on PSO is better than traditional Normalized Cross Correlation(NCC) algorithm,which can accurately estimate organization movement,and make the elastic imaging have the high quality.GPU parallel implementation of the method effectively improves the calculation speed at the same time,and also can meet the requirements of real-time ultrasound elastic imaging.

Key words: elastic imaging, cross correlation function, swarm intelligence algorithm, Particle Swarm Optimization(PSO), parallel computation, Graphic Processing Unit(GPU), Compute Unified Device Architecture(CUDA)

摘要: 为提高超声弹性成像的质量同时满足实时成像的要求,以图形处理器(GPU)为计算平台,提出一种超声弹性实时成像算法。通过描述粒子群优化估计每一个待估计点位移的过程,对该算法进行GPU并行实现,并与传统互相关算法的实验数据进行对比。仿真结果显示,该算法比传 统的互相关算法能更准确地估计组织的运动情况,使得到的弹性图具有较高的质量,同时其GPU并行实现可有效提高计算速度,满足实时超声弹性成像的要求。

关键词: 弹性成像, 互相关函数, 群智能算法, 粒子群优化, 并行计算, 图形处理单元, 统一计算设备架构

CLC Number: