Deep learning-based aberration compensation improves contrast and resolution in fluorescence microscopy
Problem
Optical aberrations reduce signal, contrast, and resolution when imaging thick samples in fluorescence microscopy.
Why it matters
Clear deep-tissue imaging is essential for biological research, but aberrations degrade it.
What the AI did
Trained neural networks on synthetically aberrated shallow-stack images to reverse aberration effects, restoring image quality without extra dose, optics, or slower acquisition.
Results
Outperformed alternative methods; restoration on par with adaptive optics across confocal, light-sheet, multi-photon, and super-resolution datasets.
Impact
Improved restored data quality across four microscopy modalities, aiding inspection and downstream analysis.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2025 |
| Work type | article |
| Authors | Min Yi Guo, Yicong Wu, Chad Michael Hobson, Yijun P. Su, Shuhao Qian, Eric E. Krueger, Ryan Patrick Christensen, Grant Kroeschell et al. |
| Citations | 54 (OpenAlex, at import time) |
| References | 62 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-024-55267-x |
| Open access | gold |
| Keywords | deep learning, aberration compensation, fluorescence microscopy, adaptive optics, image restoration |
Importance rating
real but narrower advance
Aberrations hinder fluorescence microscopy of thick samples across four modalities; a genuine capability gain for biological imaging research, but no scale of beneficiaries or deployment is stated.