Expansion Microscopy Imaging Isotropic Restoration by Unsupervised Deep Learning

Meng-Yun Wu, Da-Yu Huang, Ya-Ding Liu, Li-An Chu, Gary. Han Chang

Keywords: domain adaptation, GAN, unsupervised deep learning, expansion microscopy

TL;DR: Fluorescence light sheets microscopy imaging isotropic restoration on expansion neural microscopy by unsupervised deep learning

Abstract: The development of fluorescence light sheets and expansion microscopy (ExM) in recent years enables the visualization of detailed neural structures to help unlock the secrets of neural functioning. Deep learning techniques have then become essential tools to process the ever-increasing amount of high-quality and high-resolution images. In this study, we developed a single-scale deconvolution model for extracting multiscale deconvoluted response (MDR) from the volumes of microscopy images of neurons and generative models to translate images between the lateral and axial views. The results demonstrated that deep learning as a promising tool in approving image volume quality and comprehension of structural information of light sheet microscopy.