CS Forum: Motoki Shiga, Gifu University, Japan
CS department's public guest lecture on 'Electron Microscopic Spectral Imaging Analysis Based on Nonnegative Matrix Factorization'. The lecture is open to everyone free-of-charge.
Speaker: Motoki Shiga,
Tenure-track Assistant Professor
Affiliation: Gifu University, Japan
Host: Professor Samuel Kaski
Time: 12:15 (coffee at 12:00)
Venue: T4, CS building
Electron Microscopic Spectral Imaging Analysis Based on Nonnegative Matrix Factorization
With the development of microscopy technologies, the size of datasets to be analyzed in materials science has been rapidly increasing. Against such a large size of datasets, machine learning technologies have been becoming more important. An example of such a measurement is scanning transmission electron microscopy (STEM) with comprehensive electron energy-loss (EELS) or energy-dispersive X-ray (EDX). Spectrometers in this equipment collect a set of spectra, which called spectral imaging (SI), each from the sub-nanometer area of the sample by the fine incident electron probe consecutively scanning over the two-dimensional region. An important task on SI data analysis is to automatically identify basis spectra and spatial distributions of chemical components on a specimen. This talk introduces our developed nonnegative matrix factorization to analyze SI data and demonstrate our analysis results of real SI datasets.
Motoki Shiga is an assistant professor, department of electrical, electronic and computer engineering at Gifu University, Japan. And he is also a PRESTO researcher in Materials Informatics research area, Japan Science and Technology Agency. From 2011-2013, he was an assistant professor at Toyohashi University of Technology, and from 2006-2011, a postdoctoral researcher and an assistant professor at Kyoto University. He received his B.S. degree (Engineering) in 2001, M.S. degree (Engineering) in 2003, and Ph.D. degree (Engineering) in 2006 from Gifu University, Japan. He is working on machine learning and its application to Materials Informatics and Bioinformatics.