Deep learning-assisted low-cost autofluorescence microscopy for rapid slide-free imaging with virtual histological...

Area of Science:

• Biomedical Imaging
• Computational Pathology
• Microscopy Techniques

Background:

• Slide-free imaging offers significant improvements for histological workflows.
• Existing high-throughput methods like CHAMP microscopy provide high resolution but are costly due to specialized lasers.
• There is a need for cost-effective, high-performance imaging solutions in histology.

Purpose of the Study:

• To develop a deep learning-assisted enhanced widefield microscopy (EW-LED) framework.
• To achieve high-resolution virtual histology comparable to CHAMP microscopy.
• To reduce the cost and time associated with advanced histological imaging.

Main Methods:

• Utilized a low-cost light-emitting diode (LED) as the illumination source.
• Developed a deep learning model to enhance widefield images, targeting CHAMP microscopy outputs.
• Implemented an enhanced widefield microscopy (EW-LED) system.

Main Results:

• EW-LED achieved results comparable to CHAMP microscopy.
• The EW-LED system demonstrated an 85× cost reduction compared to CHAMP.
• Image acquisition and computation times were reduced by 36× and 17×, respectively.

Conclusions:

• EW-LED offers a highly cost-effective and efficient solution for virtual histology.
• The deep learning framework enhances widefield microscopy for improved histological analysis.
• This approach has potential applications in various imaging modalities for enhanced virtual histology.