Document Type

Article

Rights

Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence

Publication Details

First workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, 2009, pp.111-114.

Abstract

Vibrational spectroscopy (Raman and FTIR microspectroscopy) is an attractive modality for the analysis of biological samples since it provides a complete non-invasive acquisition of the biochemical fingerprint of the sample. Studies in our laboratory have applied vibrational spectroscopy to the analysis of biological function in response to external agents (chemotherapeutic drugs, ionising radiation, nanoparticles), together with studies of the pathology of tissue (skin and cervix) in health and disease. Genetic algorithms have been used to optimize spectral treatments in tandem with the analysis of the data (using generalized regression neural networks (GRNN), artificial neural networks (ANN), partial least squares modelling (PlS) and support vector machines (SVM), to optimize the complete analytical scheme and maximise the predictive capacity of the spectroscopic data. In addition we utilise variable selection techniques to focus on spectral features that provide maximal classification or regress ion of the spectroscopic data against analytical targets. This approach has yielded interesting insights into the variation of biochemical features of the biological system with its state, and has provided the means to develop further the analytical and predictive capabilities of vibrational spectroscopy in biological analysis.

DOI

https://doi.org/10.1109/WHISPERS.2009.5288989

Funder

HEA Technology Sector Research Strand 3, HEA PRTLI Cycle 4 (National Biophotonics and Imaging Platform of Ireland)


Included in

Physics Commons

Share

COinS