Abstract: Most modern analytic techniques including scanning electron microscopy, X-ray photoelectron microscopy, and Raman microscopy generate complex datasets. Data with more than 2-dimensions is often referred to as hyperspectral data. The existing analysis techniques mostly center on principal component analysis (PCA) and multivariate curve resolution (MCR). However, the interpretation of results through these methods can be complicated, especially for the uninitiated user. Moreover, results generally depend on pre-processing of the data, which can further complicate the analysis. However, the most important problem with the existing methods is a lack of capturing of the variation in the entire dataset. Here, we present a revolutionary new approach to data processing. The DAHI algorithm is a direct data dimension reduction technique that is fast owing to the structure of the algorithm. Moreover, the method captures 100% of the variance in complex datasets, and it can identify spectral components even in trace quantities. The method has been developed in the Matlab coding environment, is based on very different mathematics than PCA and MCR, and is demonstrated on data sets from energy dispersive spectrometry, time of flight secondary ion mass spectrometry and Raman microscopy. The results indicate the effectiveness of the method compared to standard PCA and MCR analysis. Finally, we demonstrate the method on a hyperspectral image data of a human face that shows the effectiveness of the data reduction in terms of capturing all the information content. Summary: Fast analysis of complex image data generated by surface imaging techniques
Ask the author a question about this poster.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Health vision 2030: An Era of Evolution and Revolution in India
Stability_testing integration program
Business Development Team
Spotlight on iPSC Reprogramming