Blind Super-resolution of Point Sources via Projected Gradient Descent
arXiv:2112.01077 · doi:10.1109/TSP.2022.3209006
Abstract
Blind super-resolution can be cast as a low rank matrix recovery problem by exploiting the inherent simplicity of the signal and the low dimensional structure of point spread functions. In this paper, we develop a simple yet efficient non-convex projected gradient descent method for this problem based on the low rank structure of the vectorized Hankel matrix associated with the target matrix. Theoretical analysis indicates that the proposed method exactly converges to the target matrix with a linear convergence rate under the similar conditions as convex approaches. Numerical results show that our approach is competitive with existing convex approaches in terms of recovery ability and efficiency.
arXiv commentarXiv admin note: text overlap with arXiv:2110.02478
References in corpus (9)
- Phase Retrieval via Wirtinger Flow: Theory and Algorithms
- User-friendly tail bounds for sums of random matrices
- Nonconvex Optimization Meets Low-Rank Matrix Factorization: An Overview
- Non-convex Optimization for Machine Learning
- Super-Resolution Delay-Doppler Estimation for OFDM Passive Radar
- Fast low-rank estimation by projected gradient descent: General statistical and algorithmic guarantees
- Guaranteed Blind Sparse Spikes Deconvolution via Lifting and Convex Optimization
- Convergence Analysis for Rectangular Matrix Completion Using Burer-Monteiro Factorization and Gradient Descent
- Mathematical Theory of Atomic Norm Denoising In Blind Two-Dimensional Super-Resolution (Extended Version)