sparse-recovery

A MATLAB library for sparse representation problems

Epistatic Net is an algorithm which allows for spectral regularization of deep neural networks to predict biological fitness functions (e.g., protein functions).

MATLAB code for reproducing the truncated Huber penalty experiments (Yang et al. , SISC 2026).)

Insense is an optimization algorithm to select a few linear measurements from a large matrix in a sparse recovery task.

Discriminative Dictionary Learning for (2D) Image Segmentation

From-scratch DSP implementations investigating FFT optimization, compressed sensing, and spectral analysis. Achieves 8x speedup and <5% reconstruction error with comprehensive testing.

Compressed Sensing reconstruction in MATLAB — OMP, L1 minimization, DCT, speech and image recovery from sub-Nyquist measurements

Fast and Efficient Data Science Techniques for COVID-19 Group Testing

This page features a series of Advanced Matrix Factorization and Decomposition.

🏁 Sparse signal recovery library written in PyCUDA.

Monte-Carlo comparison of Projection, LMMSE, LASSO and MUSIC for sparse frequency (line-spectrum) estimation - MATLAB + Python. Shows L1 sparse recovery beating classical methods in the overcomplete regime.

Presentation for the exam of Optimal Control within the PhD program in Information Engineering of the Department of Information Engineering @ University of Pisa, A.A. 2021/2022

Construction of Binary/Bipolar Compressed Sensing matrices using BCH codes

cssr is a Python package that functions as a Compressed Sensing-based super-resolution framework of (primarily) low-pass filtered and noisy ground truth signals which admit a sparse representation in some domain.