Lossless Linear Analog Compression
Abstract
We establish the fundamental limits of lossless linear analog compression by considering the recovery of random vectors from the noiseless linear measurements with measurement matrix . Specifically, for a random vector of arbitrary distribution we show that can be recovered with zero error probability from linear measurements, where denotes the lower modified Minkowski dimension and the infimum is over all sets with . This achievability statement holds for Lebesgue almost all measurement matrices . We then show that -rectifiable random vectors---a stochastic generalization of -sparse vectors---can be recovered with zero error probability from linear measurements. From classical compressed sensing theory we would expect to be necessary for successful recovery of . Surprisingly, certain classes of -rectifiable random vectors can be recovered from fewer than measurements. Imposing an additional regularity condition on the distribution of -rectifiable random vectors , we do get the expected converse result of measurements being necessary. The resulting class of random vectors appears to be new and will be referred to as -analytic random vectors.
Cite
@article{arxiv.1605.00912,
title = {Lossless Linear Analog Compression},
author = {Giovanni Alberti and Helmut Bölcskei and Camillo De Lellis and Günther Koliander and Erwin Riegler},
journal= {arXiv preprint arXiv:1605.00912},
year = {2016}
}