AVS 62nd International Symposium & Exhibition
    Applied Surface Science Thursday Sessions
       Session AS+SS-ThA

Invited Paper AS+SS-ThA6
Enhancing Chemical Contrast: Latest Trends in Hyperspectral Image Analysis

Thursday, October 22, 2015, 4:00 pm, Room 212D

Session: Advances in 2D Chemical Mapping and Data Analysis
Presenter: Barry Wise, Eigenvector Research, Inc.
Authors: B. Wise, Eigenvector Research, Inc.
W. Windig, Eigenvector Research, Inc.
Correspondent: Click to Email
Many analytical techniques that were originally developed as single sample methods have been adapted to produce multivariate aka hyperspectral images where each pixel contains an entire spectrum. These imaging techniques are used to produce “chemical maps” which elucidate the arrangement and chemical makeup of areas on the surface. Imaging techniques produce an avalanche of data which is often underutilized. Multivariate statistical methods are increasingly being used to capture the information in these large data sets, condense it down to a manageable level, and improve the signal to noise ratio. This serves to enhance the chemical contrast in the images. This talk surveys methods for hyperspectral image analysis, discusses their pros and cons, and gives some examples demonstrating what can be accomplished. Methods considered include Principal Components Analysis (PCA), Multivariate Curve Resolution (MCR) with contrast constraints, Independent Components Analysis (ICA), Maximal Autocorrelation Factors (MAF) and sample clustering techniques.