计算机应用 ›› 2011, Vol. 31 ›› Issue (08): 2123-2125.DOI: 10.3724/SP.J.1087.2011.02123

• 人工智能 • 上一篇    下一篇

迭代硬阈值压缩感知重构算法——IIHT

张宗念1,李金徽2,黄仁泰3   

  1. 1. 东莞理工学院 电子工程学院,广东 东莞523808
    2. 东莞理工学院 网络中心,广东 东莞523808
    3. 东莞理工学院 计算机学院,广东 东莞523808
  • 收稿日期:2011-01-17 修回日期:2011-02-24 发布日期:2011-08-01 出版日期:2011-08-01
  • 通讯作者: 张宗念
  • 作者简介:张宗念(1963-),男,河北深州人,副教授,博士,主要研究方向:压缩感知、图像分析与处理;李金徽(1980-),男,辽宁沈阳人,工程师,主要研究方向:分布式计算机网络;黄仁泰(1964-),男,广东东莞人,副教授,主要研究方向:嵌入式系统设计。
  • 基金资助:

    广东省自然科学基金资助项目(9151170003000017)

IIHT: New improved iterative hard thresholding algorithm for compressive sensing

Zong-nian ZHANG1,Jin-hui LI2,Ren-tai HUANG3   

  1. 1. School of Electronic Engineering, Dongguan University of Technology, Dongguan Guangdong 523808, China
    2. Network Center, Dongguan University of Technology, Dongguan Guangdong 523808, China
    3. School of Computer Science, Dongguan University of Technology, Dongguan Guangdong 523808, China
  • Received:2011-01-17 Revised:2011-02-24 Online:2011-08-01 Published:2011-08-01
  • Contact: Zong-nian ZHANG

摘要: 研究了压缩感知信号重构算法的理论,针对迭代硬阈值(IHT)重构算法对测量矩阵的过分依赖、计算复杂度高、运算时间长的缺点,通过修订迭代硬阈值重构算法的代价函数和自适应地调整迭代步长的选取原则,设计了一种迭代硬阈值重构算法——IIHT。IIHT算法显著提高了信号精确重构的概率,降低了算法的计算复杂度,进一步减少了算法的运算时间,加快了算法的收敛速度。

关键词: 迭代, 硬阈值, 压缩感知

Abstract: To overcome the shortcomings of the overdependence on the measurement matrix, the high computation complexity, the long computation time of the Iterative Hard Thresholding (IHT) algorithm, a new improved iterative hard thresholding (IIHT) algorithm was proposed by studying the theory of signal reconstruction for compressive sensing. It improved the cost function and the selection method of step size for the IHT algorithm. The simulation results show that the proposed algorithm increases the probability of recovery and the speed of convergence and reduces the computational complexity and time.

Key words: iteration, hard thresholding, compressive sensing

中图分类号: