Robust AI-aided Imaging Models with Unpaired Data

Abstract

The observations in practical imaging systems always contain complex noise such that classical approaches are difficult to obtain satisfactory results. In recent years, deep neural networks directly learned a map between the noisy and clean images based on the training on paired data. Despite its promising results in various tasks, collecting the training data is difficult and time-consuming in practice. In this talk, in the unpaired data regime, we will discuss our recent progress for building AI-aided robust models and their applications in image processing. Leveraging the Bayesian inference framework, our model combines classical mathematical modelling and deep neural networks to improve interpretability. Experimental results on various real datasets validate the advantages of the proposed methods. Finally, I will report the recent progresses on solving the preferred orientation problems in cyroEM using the developed tools.

Reference:

Addressing preferred orientation in single-particle cryo-EM...

The single-particle cryo-EM field faces the persistent challenge of preferred orientation, lacking general computational solutions. We introduce cryoPROS, an AI-based approach designed to address...

https://arxiv.org/abs/2309.14954

Learn from Unpaired Data for Image Restoration: A Variational...

Collecting paired training data is difficult in practice, but the unpaired samples broadly exist. Current approaches aim at generating synthesized training data from unpaired samples by exploring...

https://arxiv.org/abs/2204.10090

Procedure of cyroEM:

One mentioned data set:

EMPIAR-10096 Untilted Single-Particle CryoEM of Highly Preferred Orientated Influenza Hemagglutinin Trimer

EMPIAR, the Electron Microscopy Public Image Archive centered at EMBL-EBI, is a public resource for raw electron microscopy images related to EMDB, contains micrographs, particle sets and tilt-series.

https://www.ebi.ac.uk/empiar/EMPIAR-10096/