Performance of multiple-plane phase retrieval using intensity patterns at various regions near the image plane

Abstract

Phase information is typically lost when using conventional imaging devices that capture only the intensity. To recover the phase information, multiple techniques are employed, including interferometric and iterative methods. Single beam multiple intensity reconstruction (SBMIR) is an iterative phase retrieval algorithm that uses a stack of images captured at multiple distances. Most techniques use only images captured in the direction of forward propagation. In this purely numerical study, conventional SBMIR is applied to simulated subsets of intensity patterns captured at various distances before and after the object plane, and some subsets that include the measurement at the focused image plane. The effect of using different subsets on the convergence speed of SBMIR is quantified using root-mean-square error (RMSE). Results show that for the object wave used, the distance of the image stack from the focused image plane impacts the convergence speed. The effect of the inclusion of the focused image plane is also investigated for phase-type and amplitude-type objects. This study may have applications in the optimization of techniques that utilize intensity
maps at multiple propagation distances.