LAUSR

Abstract Multi-distance phase retrieval enables a full wavefront recovery with a set of defocused intensity patterns. However, the convergence speed of multi-distance phase retrieval is restricted by a stagnant issue, which leads to a low imaging contrast for visual observation. We propose an acceleration scheme to solve this problem. A shrink-wrap constraint is imposed to obtain an efficient support region. In support region, a weighted feedback strategy is used to speed up the convergence. In non-support region, an attenuated term is introduced to eliminate redundant data. With this combined constraint, imaging contrast and convergence speed are both enhanced observably. Simulations and experiments demonstrate the validity of this acceleration scheme in a lensfree imaging system. The proposed method can both reduce the workload of experimental data acquisition and computational load.