Blind super-resolution of point sources via fast iterative hard thresholding

2021/09/23 by Zhu, Zengying, Chen, Jinchi, Gao, Weiguo
#FOS: Computer and information sciences #Information Theory (cs.IT)

paper · doi:10.48550/arxiv.2109.11414

Abstract

In this work, we develop a provable fast algorithm for blind super-resolution based on the low rank structure of vectorized Hankel matrix associated with the target matrix. Theoretical results show that the proposed method converges to the ground truth with linear convergence rate. Numerical experiments are also conducted to illustrate the linear convergence and effectiveness of the proposed approach.

Related

• Blind Super-resolution via Projected Gradient Descent
• Simultaneous Blind Demixing and Super-resolution via Vectorized Hankel Lift
• Simpler Gradient Methods for Blind Super-Resolution with Lower Iteration Complexity
• Vectorized Hankel Lift: A Convex Approach for Blind Super-Resolution of Point Sources
• Fast and Provable Simultaneous Blind Super-Resolution and Demixing for Point Source Signals: Scaled Gradient Descent without Regularization