We will discuss a fast, randomized approximate nearest neighbor algorithm (joint work with Andrei Osipov and Vladimir Rokhlin) and discuss applications to imaging problems, where nearest neighbor calculations are often a computational bottleneck. In the first part of the lecture I will discuss the algorithm. Part two of the lecture will be given by Andrei Osipov, and he will demonstrate the algorithm on imaging data. Part three of the lecture will be given by Zhizhen Zhao. She will describe some projects in Amit Singer's group (Princeton) where nearest neighbor problems appear in Cryo-EM, and where our algorithm is used.
Andrei Osipov and Zhizhen Zhao, Co-presenters
Copyright. All Rights Reserved.