A noise-robust multi-intensity phase retrieval method based on structural patch decomposition

Multi-image iterative parallel phase retrieval methods are widely used in measurement and imaging, due to their simplicity and effectiveness. However, the retrieved image will have foggy diffuse noise, since some unknown cross-talk factors exist in iterative calculation. In this paper, a new updated method is embedded into the calculation process of phase retrieval. The iterative strategy of phase retrieval is updated to enhance robustness for noisy images. Experimental and simulated results have demonstrated that the improved algorithm has faster convergence speed and higher anti-noise performance. (paper)