Calibration-free quantitative phase imaging using data-driven aberration modeling.

Area of Science:

• Optical Physics
• Biomedical Imaging
• Machine Learning

Background:

• Quantitative Phase Imaging (QPI) is crucial for label-free cell imaging.
• Optical aberrations degrade QPI image quality and quantitative accuracy.
• Existing aberration correction methods require complex setups or additional measurements.

Purpose of the Study:

• To develop a data-driven method for aberration correction in QPI.
• To eliminate the need for background regions or extra calibration measurements.
• To enable single-shot aberration correction using deep learning.

Main Methods:

• A U-net-based deep neural network was employed.
• The network learned a translation from aberrated to aberration-corrected phase fields.
• The method was validated on 2D and 3D QPI data.

Main Results:

• A single-shot aberration-corrected field image was successfully generated.
• The deep learning approach demonstrated high fidelity and stability.
• Performance was benchmarked against conventional background subtraction methods.

Conclusions:

• Deep learning offers an efficient and accurate solution for QPI aberration correction.
• This method simplifies the QPI workflow by removing calibration steps.
• The technique is robust for imaging diverse biological samples and calibration standards.