Deep learning for data synthesis in fluorescence microscopy

Abstract number
192
Presentation Form
Invited
Corresponding Email
[email protected]
Session
Session 5 - Super-resolution and Nanoscale Imaging
Authors
Susan Cox (1)
Affiliations
1. King's College London
Keywords
Abstract text

In cryo-EM synthesis of information from multiple views is well established as a method for inferring molecular structure. We have developed a way to extend this technique to the mesoscale using fluorescence microscopy images of multiple views of a single structure as inputs, with deep learning accelerating the fit of the angle of each view. Using this approach we are able to perform completely free fits and reproduce nanoscale structure.

References