计算机技术与发展2017，Vol.27Issue(10)：69-73,5.DOI:10.3969/j.issn.1673-629X.2017.10.015
改进变步长快速迭代收缩阈值算法
An Improved Fast Iterative Shrinkage-thresholding Algorithm with Variable Stepsize
摘要
Abstract
Image restoration is an important research in image processing and solving of it involves large databases and many unknowns. The forward-backward splitting method provides a simple,practicable solution to solve the kinds of data optimization with high dimen-sion. The fast iterative shrinkage-thresholding algorithm joins the global acceleration operator and improve its convergence rate based on the forward-backward algorithm. However,the fixed step-size limits its speed of the convergence in minimal optimization solving. To ad-dress that,an improved algorithm with variable stepsize by using Barzilai-Borwein ( BB) operator is proposed which updates the step size with the iterative information of the first two steps in each iteration to accelerate its convergence. Then it' s applied to image denosing and compressed sensing. The experimental results demonstrate that it is more efficient than the original fast iterative shrinkage-threshold algo-rithm,not only in the efficiency but also in the signal to noise ratio of signal restoration.关键词
快速迭代收缩阈值算法/Barzilai-Borwein算子/全变分模型/压缩感知/图像去噪Key words
fast iterative shrinkage-threshold algorithm/Barzilai-Borwein operator/total variation model/compressed sensing/image de-nosing分类
信息技术与安全科学引用本文复制引用
陈少利,杨敏..改进变步长快速迭代收缩阈值算法[J].计算机技术与发展,2017,27(10):69-73,5.基金项目
国家自然科学基金资助项目(61271234) （61271234）
