AI-enhanced visual pattern recognition for diagnostic electron microscopy using the example of viruses

Short description of the project

Diagnostic electron microscopy (EM) visualizes human tissue from the millimeter to the nanometer scale and is thereby able to detect pathological changes and foreign matter, such as viruses, bacteria or parasites. The recognition of objects in diagnostic EM is performed by human experts using visual object recognition while performing microscopy. This procedure allows a fast inspection of diagnostic samples but needs considerable expertise and long-time training. The aim is to develop Deep Learning methods that support diagnosis by automatically detecting and classifying virus particles.

The project addresses the following “Essential Public Health Functions” (EPHS):

• #1: Assess and monitor population health status, factors that influence health, and community needs and assets.
  On a large set of unknown data (EM images) diagnostics could be performed in very short time with a deep learning-based software solution, which would support monitoring health status.
• #2: Investigate, diagnose, and address health problems and hazards affecting the population.
  Diagnostics on EM images with a deep learning-based software solution could complement other diagnostic methods.
• #9: Improve and innovate public health functions through ongoing evaluation, research, and continuous quality improvement.
  Quality of diagnostic electron microscopy might be improved (computational precision in recognizing details, no human bias)

to the top