Unsupervised inter-domain transformation for virtually stained high-resolution mid-infrared photoacoustic microscopy using explainable deep learning
2024

Transforming Low-Resolution Mid-Infrared Images into High-Resolution Stained Images Using Deep Learning

Sample size: 16 publication 10 minutes Evidence: high

Author Information

Author(s): Park Eunwoo, Misra Sampa, Hwang Dong Gyu, Yoon Chiho, Ahn Joongho, Kim Donggyu, Jang Jinah, Kim Chulhong

Primary Institution: Pohang University of Science and Technology (POSTECH), Pohang, Republic of Korea

Hypothesis

Can explainable deep learning improve the resolution and staining of mid-infrared photoacoustic microscopy images?

Conclusion

The study successfully demonstrates a method to transform low-resolution mid-infrared images into high-resolution, virtually stained images using explainable deep learning.

Supporting Evidence

  • The explainable deep learning framework improved the stability and reliability of image transformations.
  • Label-free imaging of human cardiac fibroblasts was achieved with enhanced resolution.
  • The method allows for quantitative analysis of cell growth and protein expression.

Takeaway

This study shows how computers can help make blurry pictures of cells look clear and colorful without using any dyes.

Methodology

The study used an explainable deep learning framework to transform low-resolution mid-infrared images into high-resolution images, validated by comparing with confocal fluorescence microscopy images.

Limitations

The image resolution of mid-infrared photoacoustic microscopy is still not sufficient to distinguish detailed cellular structures.

Participant Demographics

Human cardiac fibroblasts were used for imaging.

Statistical Information

P-Value

1.3×10−16

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1038/s41467-024-55262-2

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication