Sorry, you need to enable JavaScript to visit this website.

Data-Driven Algorithms for Gaussian Measurement Matrix Design in Compressive Sensing

Citation Author(s):
Yang Sun, Jonathan Scarlett
Submitted by:
Yang Sun
Last updated:
6 May 2022 - 5:55am
Document Type:
Presentation Slides
Document Year:
2022
Event:
Presenters:
Yang Sun
Paper Code:
SPTM-5.3

Abstract

In this paper, we provide two data-driven algorithms for learning compressive sensing measurement matrices with Gaussian entries. In contrast to the ubiquitous i.i.d.~Gaussian design, we associate different variances with different signal entries, so that we may utilize training data to focus more energy on the ``most important'' parts of the signal. Our first algorithm is based on simple variance-proportional sampling (i.e., place more energy at locations where the signal tends to vary more), and our second overcomes limitations of the first by iteratively up-weighing and down-weighing the variance values according to reconstructions performed on the training signals. Our algorithms enjoy the advantages of being simple and versatile, in the sense of being compatible with a diverse range of signal priors and/or decoding rules. We experimentally demonstrate the effectiveness of our algorithms under both generative priors with gradient-based recovery and sparse priors with l1-minimization based recovery.

up
0 users have voted:

Files

The slides of the paper Data-Driven Algorithms for Gaussian Measurement Matrix Design in Compressive Sensing

(54)

The poster of the paper Data-Driven Algorithms for Gaussian Measurement Matrix Design in Compressive Sensing

(52)