What is it about?

Microscopes are essential for studying cells and identifying diseases. But high-quality microscopes are costly. This is because it uses multiples lenses that needs mounting and testing. One way to reduce its cost is to use a single lens, the ones used in cellphone cameras. But these lenses cannot focus different colors present in visible light at the same spot. As a result, magnified images of colored samples appear blurred. In this paper, the authors proposed a method for solving this issue. They shone a light of single color over a tissue sample for three different colors, namely red (R), blue (B), and green (G) light. This allowed the single lens to focus their images clearly. At the same time, an image sensor recorded these images. The authors then used these single-color images as input to a deep learning network. They then trained this network to color these images based on colored (RGB) image samples.

Featured Image

Why is it important?

To study cells under a microscope, they often need to be colored to distinguish them. This is done with coloring dyes. The microscope must correctly reflect the color of dyed samples. But this needs the use of multiple lenses, making the microscope costly. This study proposes a method based on machine learning that transforms grayscale image into color images. KEY TAKEAWAY: While single lenses have difficulty focusing different colors, they can take detailed single-color images. By adding color to these images using machine learning, this study could make low-cost microscopes a reality.

Read the Original

This page is a summary of: Deep learning virtual colorization overcoming chromatic aberrations in singlet lens microscopy, APL Photonics, March 2021, American Institute of Physics, DOI: 10.1063/5.0039206.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page