LED-based temporal variant noise model for Fourier ptychographic microscopy.

Area of Science:

• Optics and Imaging Science
• Microscopy Techniques
• Image Processing

Background:

• Fourier ptychographic microscopy (FPM) reconstructs high-resolution images from low-resolution inputs.
• FPM is susceptible to various noise sources, particularly in large-angle and dark-field imaging.
• Effective noise reduction is crucial for accurate FPM image reconstruction.

Purpose of the Study:

• To develop an adaptive denoising algorithm for Fourier ptychographic microscopy.
• To address noise issues in FPM, especially concerning temporal variants and LED illumination.
• To improve the quality and detail recovery of reconstructed high-resolution images.

Main Methods:

• Proposed an adaptive denoising algorithm utilizing a LED-based temporal variant noise model.
• Used blank slide samples to establish a reference noise value.
• Developed a statistical model linking temporal noise to LED spatial location.
• Applied adaptive Gaussian denoising based on the noise model.

Main Results:

• The proposed algorithm effectively suppresses noise in FPM images.
• Enhanced recovery of image details and increased image contrast were observed.
• Superior visual effects and objective evaluation results compared to other methods.
• Demonstrated improved high-resolution image reconstruction quality.

Conclusions:

• The adaptive denoising algorithm significantly improves FPM image quality.
• The LED-based temporal variant noise model accurately characterizes and mitigates noise.
• The method offers a robust solution for noise reduction in FPM, leading to better reconstructions.