Abstract:Fourier phase retrieval (PR) is a severely ill-posed inverse problem that arises in various applications. To guarantee a unique solution and relieve the dependence on the initialization, background information can be exploited as a structural prior. However, the requirement for the background information may be challenging when moving to high-resolution imaging. At the same time, the previously proposed projected gradient descent (PGD) method also demands much background information. In this paper, we present an improved theoretical result about the demand for the background information, along with two Douglas Rachford (DR) based methods. Analytically, we demonstrate that the background information required to ensure a unique solution can be decreased by nearly for the 2-D signals compared to the 1-D signals. By generalizing the results into d-dimension, we show that the length of the background information more than folds of the signal is sufficient to ensure uniqueness. At the same time, we also analyze the stability and robustness of the model when the measurements and background information are corrupted by noise. Furthermore, two methods called Background Douglas Rachford (BDR) and Convex Background Douglas Rachford (CBDR) are proposed. BDR, which is a kind of non-convex method, is proven to have the local R-linear convergence rate under mild assumptions. Instead, the CBDR method uses the techniques of convexification and can be proven to have a global convergence guarantee as long as the background information is sufficient. To support this, a new property called F-RIP is established. We test the performance of the proposed methods through simulations as well as real experimental measurements, and demonstrate that they achieve a higher recovery rate with less background information compared to the PGD method.

On Global Convergence of Gradient Descent Algorithms for Generalized Phase Retrieval Problem.

Riemannian Optimization for Phase Retrieval from Masked Fourier Measurements

Convolutional Phase Retrieval via Gradient Descent

PhaseMax: Stable Guarantees from Noisy Sub-Gaussian Measurements

Global convergence of gradient descent for phase retrieval

Convergence of Projected Subgradient Method with Sparse or Low-Rank Constraints

Convergence Analysis of Projected Gradient Descent for Schatten-p Nonconvex Matrix Recovery

Phase Retrieval with PhaseLift Algorithm

Nearly optimal bounds for the global geometric landscape of phase retrieval

Globally Convergent Levenberg-Marquardt Method for Phase Retrieval.

Strong Convexity of Affine Phase Retrieval

Linear Convergence of an Iterative Phase Retrieval Algorithm with Data Reuse

A Fast and Provable Algorithm for Sparse Phase Retrieval

Phase Retrieval Via Reweighted Amplitude Flow

Robust Gradient Descent for Phase Retrieval

Over-Parameterized Network Solves Phase Retrieval Effectively

Low Complexity Regularized Phase Retrieval

Phase Retrieval via Linear Programming: Fundamental Limits and Algorithmic Improvements

Convergence Analysis of Gradient Algorithms on Riemannian Manifolds Without Curvature Constraints and Application to Riemannian Mass

Convergence Analysis on a Fast Iterative Phase Retrieval Algorithm Without Independence Assumption

Phase Retrieval With Background Information: Decreased References and Efficient Methods