Real-time tissue characterization on the basis of in vivo Raman spectra

The application of in vivo Raman spectroscopy for clinical diagnosis demands dedicated software that can perform the necessary signal processing and subsequent (multivariate) data analysis, enabling clinically relevant parameters to be extracted and made available in real time. Here we describe the design and implementation of a software package that allows for real-time signal processing and data analysis of Raman spectra. The design is based on automatic data exchange between Grams, a spectroscopic data acquisition and analysis program, and Matlab, a program designed for array-based calculations. The data analysis software has a modular design providing great flexibility in developing custom data analysis routines for different applications. The implementation is illustrated by a computationally demanding application for the classification of skin spectra using principal component analysis and linear discriminant analysis. Copyright (C) 2002 John Wiley Sons, Ltd.