Microscopy with Deep Learning – Enhancing super-resolution microscopy with deep learning

Context :

Super-resolution microscopy has become a crucial technology in life sciences and has fueled important discoveries in cell biology and neurobiology. Single molecule localization microscopy (SMLM) is one of the most popular forms of super-resolution microscopy, but still faces major bottlenecks such as poor temporal resolution and limited spectral content. The recent reemergence of deep learning has led to a number of methods that overcome some of these limitations. Specifically, we recently developed ANNA-PALM, a computational method that uses deep learning to accelerate SMLM1, thereby enabling high-throughput super-resolution imaging. 

Objectives :

In this project, we aim to extend ANNA-PALM for (i) live cell super-resolution imaging, (ii) three-dimensional imaging without scanning, (iii) multicolor imaging with single dyes, (iv) end-to-end reconstruction from raw low resolution images, and (v) providing tools to better safeguard against reconstruction artifacts.

Expected results :

The methods that will be developed in this project will enable considerable advances in our ability to image biological structures at high resolution. As these methods are largely generic and purely computational, they will be easily transferrable to many existing microscopy systems and therefore have high economic potential, which our partnership with Abbelight will allow us to achieve.