Fourier ptychography microscopy for digital pathology.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Medical Diagnostics

Background:

• Fourier ptychography microscopy (FPM) has rapidly advanced since 2013, offering high resolution and wide field-of-view imaging.
• Its adaptability across optical wavelengths enables diverse medical applications, including digital pathology, radiology, and UV imaging.
• Computational algorithms are pivotal to FPM's progress and performance.

Purpose of the Study:

• To review the fundamental physical and computational concepts driving FPM advancements in digital pathology.
• To evaluate the impact of computational algorithms, from Gerchberg-Saxton to deep learning, on FPM performance.
• To discuss the implications of FPM algorithms for automated diagnosis in digital pathology and future research.

Main Methods:

• Review of foundational physical principles in Fourier ptychography microscopy.
• Analysis of computational algorithm evolution, including phase retrieval and deep learning.
• Exploration of FPM applications in digital pathology and medical imaging.

Main Results:

• Significant improvements in FPM resolution and field-of-view are attributed to computational algorithm development.
• Early algorithms like Gerchberg-Saxton have been surpassed by advanced phase-retrieval and deep-learning techniques.
• These algorithmic advancements enhance FPM's utility for automated diagnosis in digital pathology.

Conclusions:

• Computational algorithms are the primary drivers of Fourier ptychography microscopy's success in medical imaging.
• Deep learning and advanced phase-retrieval methods represent the forefront of FPM algorithm development.
• FPM, powered by sophisticated algorithms, holds significant promise for the future of automated digital pathology and medical diagnostics.