Abstract: “Lensless holographic imaging is challenged by the twin-image artifact due to the missing phase and the aliasing effect due to the undersampled measurement. Therefore, phase retrieval and pixel super-resolution (PSR) techniques serve as the essential ingredients for high-fidelity holographic imaging. In this work, we combine the two in a unified framework by formulating the PSR phase retrieval as a non-convex feasibility problem. An adaptive smoothing strategy for escaping local minima is introduced. Numerical and experimental results are presented and discussed. The proposed framework can be generalized to various physical settings, and is compatible with the state-of-the-art iterative projection algorithms.”
