Quantitative phase imaging in dual-wavelength interferometry using a single wavelength illumination and deep...

Area of Science:

• Optics and Photonics
• Artificial Intelligence in Imaging
• Metrology

Background:

• Quantitative phase imaging (QPI) is crucial for analyzing transparent samples.
• Phase-shifting interferometry (PSI) faces limitations in measurement range and complexity, often requiring multiple wavelengths.
• Single-wavelength interferometry (SWI) offers simplicity but has a restricted measurement range.

Purpose of the Study:

• To develop a novel deep learning-based quantitative phase imaging method.
• To achieve dual-wavelength phase recovery using only single-wavelength illumination.
• To overcome the inherent measurement range limitations of SWI.

Main Methods:

• Utilizing a conditional generative adversarial network (CGAN) for phase recovery.
• Generating synthetic wavelength interferograms and wrapped phases from a single recorded interferogram.
• Validating the proposed method through numerical simulations and experimental procedures.

Main Results:

• The deep learning approach successfully recovered phases at synthetic wavelengths.
• The measurement range of SWI was significantly improved.
• The method avoids the complexities associated with simultaneous dual-wavelength illumination.

Conclusions:

• The proposed deep learning-based QPI method effectively extends the measurement range of SWI.
• This technique offers a simplified solution for phase unwrapping and range limitations in PSI.
• The findings present a promising advancement for optical metrology and imaging applications.