Core Concepts
Introducing a novel stochastic ADMM algorithm for large-scale ptychography with weighted total variation to enhance image reconstruction quality.
Abstract
The content introduces a stochastic ADMM algorithm for large-scale ptychography, addressing challenges in image reconstruction. It discusses the importance of incorporating total variation regularization and presents numerical results comparing different algorithms.
Introduction
Ptychography combines diffractive imaging and microscopy.
Various algorithms exist for phase retrieval in ptychography.
Mathematical Model
Describes notations and formulations used in the context of ptychography.
Introduces AITV-regularized variants to improve image recovery.
Convergence Analysis
Discusses the convergence of the proposed algorithm under specific assumptions.
Establishes conditions for subsequential convergence to KKT points.
Numerical Results
Evaluates the performance of the algorithm on complex images corrupted by Gaussian or Poisson noise.
Compares results with other algorithms like Douglas-Rachford splitting and rPIE.
Stats
"Large-scale ptychography presents challenges in memory usage and computational cost."
"AITV has shown better performance than TV in various image processing tasks."
Quotes
"Ptychography is a prevalent imaging technique combining diffractive imaging and microscopy."
"Total variation regularization improves robustness when measurements are corrupted by noise."