Structure-aware phase aberration compensation for digital holographic microscopy

Digital holographic microscopy (DHM) enables label-free imaging and quantitative characterization of transparent or weakly absorbing samples by measuring phase delay. However, optical aberrations introduce wavefront distortions, thereby degrading the accuracy and contrast of phase reconstruction. To address this problem, we propose a structure-aware phase aberration compensation method for DHM, which adaptively estimates and removes system aberrations from a single phase map. The proposed method uses multiscale phase gradients, local curvature, and local variance to characterize structural saliency in the sample and construct a continuous background confidence map. By further incorporating robust residual soft clipping and graph-neighborhood regularization, the method suppresses interference from sample edges, strong phase variations, noise, and local reconstruction artifacts during aberration estimation, enabling stable aberration compensation without manual background annotation. Simulation and experimental results demonstrate that the method is robust to aberrations of different orders and can effectively estimate and correct system aberrations, yielding high-quality quantitative phase images. This study provides a robust and effective solution for quantitative phase recovery in DHM.