DeepPD: Joint Phase and Object Estimation from Phase Diversity with Neural Calibration of a Deformable Mirror
Journal:
arXiv
Published Date:
Apr 19, 2025
Abstract
Sample-induced aberrations and optical imperfections limit the resolution of
fluorescence microscopy. Phase diversity is a powerful technique that leverages
complementary phase information in sequentially acquired images with
deliberately introduced aberrations--the phase diversities--to enable phase and
object reconstruction and restore diffraction-limited resolution. These phase
diversities are typically introduced into the optical path via a deformable
mirror. Existing phase-diversity-based methods are limited to Zernike modes,
require large numbers of diversity images, or depend on accurate mirror
calibration--which are all suboptimal. We present DeepPD, a deep learning-based
framework that combines neural representations of the object and wavefront with
a learned model of the deformable mirror to jointly estimate both object and
phase from only five images. DeepPD improves robustness and reconstruction
quality over previous approaches, even under severe aberrations. We demonstrate
its performance on calibration targets and biological samples, including
immunolabeled myosin in fixed PtK2 cells.
