时间：12月22日下午3:00-4:00

地点：勤园21-304

摘要：We are concerned with the problem of recovering sparse vectors by orthogonal matching pursuit (OMP). Suppose that the overcomplete dictionary A consists of pairs of orthonormal bases. Let $\mu(A)$ denote the mutual coherence of A. We show that OMP can exactly recover every s sparse vector x from y = Ax in s iterations if and only if $s<\frac{\sqrt{2}-0.5}{\mu(A)}.$ The result gives a positive answer to an open question asked by Bruckstein, Donoho and Elad in 2009. This work is jointed with Prof. Song Li and Chenyun Yu.

报告人：沈益 教授

报告人简介：沈益，浙江理工大学数学科学系教授，2009年毕业于浙江大学数学系获应用数学博士学位。从事小波分析及应用、压缩感知等相关领域的研究。研究内容为信号处理，图像处理中的数学问题与方法。主持国家自然科学基金优秀青年基金项目，面上项目，浙江省杰出青年基金项目等省部级项目。在《Applied and Computational Harmonic Analysis》、《IEEE Transaction on Information Theory》和《IEEE Transactions on Signal Processing》等期刊发表SCI论文20余篇。