Dimensionality reduction technique based phase aberration compensation and spurious fringe removal in off-axis digital holographic microscopy

A dimensionality reduction technique based on singular value decomposition (SVD) is proposed for the aberration and spurious fringe removal from the phase measurement in off-axis digital holographic microscopy. The SVD of complex-valued virtual image obtained from numerically reconstructed hologram is computed. The phase aberration and spurious fringe compensated phase estimate is obtained by appropriate selection of singular values to reconstruct the denoised phase image. In order to remove the artifacts created along the x and y axis due to the SVD, the filtering procedure is implemented by rotating the phase image a fixed number of times. The effective denoised image is obtained by the weighted combination of denoised image evaluated for each rotation. Simulation and experimental studies are conducted to demonstrate the practical applicability of the proposed method. •We proposed phase aberration removal using SVD-based dimensionality reduction.•Optimal singular values are selected to achieve an aberration-free image.•Avoids 2D phase unwrapping and least-square fitting as required in PCA-based methods.•Our method is noise robust and compensates for quadratic and high-order aberration.