This is the code to the paper: Gradient-Based Learning of Discrete Structured Measurement Operators for Signal Recovery by:

Jonathan Sauder (EPFL)

Martin Genzel (Utrecht University)

Peter Jung (TU Berlin)

An example for how to use this codebase for the Single-Pixel Imaging experiments from the paper is provided in main_singlepixel.py.

Please write us an email in case of questions.

Requirements: torch, torchvision, scipy, numpy, matplotlib, pandas, pywt