In this talk, the invariants of the cryo-EM problem will be introduced. We will start by discussing the group invariants of the simplified model of multireference alignment and show how these invariants can also be used for the 2D classification problem in cryo-EM. Then, it will be shown how by similar principles we construct an ab initio model from the cryo-EM images, without estimating the viewing directions, using low-order moments.

This presentation is part of Minisymposium “MS21 - Recent mathematical advances in phase retrieval and computational imaging (2 parts)”
organized by: Mahdi Soltanolkotabi (University of Southern California) , Tamir Bendory (Princeton University) .

Authors:

Tamir Bendory (Princeton University)

Keywords:

cryo-em, group invariants, image reconstruction, inverse problems, multireference alignment, nonlinear optimization