Learning the Cellular Architecture from its Optical Properties

A video presentation from Dr. Shalin Mehta at the AI Microscopy Symposium on ML and DL technologies

Analysis of anatomy and axon orientation of an adult mouse brain tissue with QLIPP Learning_the_Cellular_Architecture_from_its_Optical_Properties_teaser.jpg

In the last 3 years, microscopists have started to use "AI based" solutions for a wide range of applications, including image acquisition optimization (smart microscopy), object classification, image classification, segmentation, restoration, super resolution and virtual staining.
The 6th edition of the AI Microscopy Symposium offered a unique forum for presenting and discussing the latest AI-based technologies and tools in the field of microscopy and biomedical imaging. Additionally, the symposium highlighted scientific breakthroughs which are enabled by machine learning or deep learning technologies.

Image: Analysis of anatomy and axon orientation of an adult mouse brain tissue with QLIPP

Speaker: Shalin Mehta, Ph.D.

Shalin received his PhD at the National University in Singapore in optics. After this graduate work, he was a staff scientist at the University of Chicago and a Human Frontier Science Program Fellow at the Marine Biological Laboratory in Woods Hole. Now Shalin is a platform leader in computational microscopy at the Chan Zuckerberg Biohub.

At the CZ Biohub, his team develops imaging and computational technologies to probe the physical properties of biological systems with more precision and efficiency. It is designed to work on multiple scales from organelles to tissues and integrates research from many fields including optics, machine learning and inverse algorithms.

More about his work: https://www.czbiohub.org/comp-micro/

Key Learnings

  • Quantitative label free cell imaging with phase and polarization can separate cellular structures through physical properties while providing a gentle environment for cells   
  • 2.5D UNet based deep learning models and quantitative label free imaging are computationally more efficient and almost as accurate as 3D UNets in virtual staining applications
  • Cellular morphological states and transitions can be extracted from quantitative label free images using DynaMorph, a deep-learning framework

Register to view the video presentation

By clicking on the SUBMIT button, I confirm that I have reviewed and agree with Leica Microsystems GmbH Terms of Use and their Privacy Policy. I also understand my privacy choices as they pertain to my personal data, as detailed in the aforementioned Privacy Policy under ‘’Your Privacy Choices’’.

Related Articles

Interested to know more?

Talk to our experts. We are happy to answer all your questions and concerns.

Contact Us

Do you prefer personal consulting?