Deterministic-Iterative Integrated Phase Retrieval

Abstract

In multi-plane phase retrieval imaging, the accuracy and efficiency of phase retrieval algorithm are usually mutually restrictive. Specifically, deterministic algorithms struggle to achieve sufficient accuracy, while iterative algorithms consume excessive time, thereby limiting their practical application. To address this issue, we propose a deterministic-iterative integrated phase retrieval algorithm, that is, an approximate phase, which could be quickly obtained by the deterministic algorithm, is imported as an initial value into the iterative algorithm to retrieve more accurate result efficiently. To demonstrate the effectiveness of this algorithm, we simulate its performance under various iterations and diffraction distances. Additionally, experiments are conducted using a pure-phase USAF1951 target and fixed mouse fibroblasts to verify its feasibility, high accuracy, and rapid iterative convergence speed. Integrating the deterministic and iterative algorithms, this method offers a novel approach for enhancing phase retrieval.