Information Theoretic Limits for Phase Retrieval with Subsampled Haar
Sensing Matrices
We study information theoretic limits of recovering an unknown dimensional, complex signal vector with unit norm from magnitude-only measurements of the form , where is the sensing matrix. This is known as the Phase Retrieval problem and models practical imaging systems where measuring the phase of the observations is difficult. Since in a number of applications, the sensing matrix has orthogonal columns, we model the sensing matrix as a subsampled Haar matrix formed by picking columns of a uniformly random unitary matrix. We study this problem in the high dimensional asymptotic regime, where , while with being a fixed number, and show that if , then any estimator is asymptotically orthogonal to the true signal vector . This lower bound is sharp since when $m > (2+o_n(1)) \cdot n $, estimators that achieve a non trivial asymptotic correlation with the signal vector are known from previous works.
View on arXiv