Figure 1
From: Similarity maps and hierarchical clustering for annotating FT-IR spectral images
Principle of FT-IR microscopy. (A) At each pixel (indicated by circles), the infrared absorbance spectrum is measured, reflecting the biochemical status of the sample at the corresponding position. (B) Average spectra for different tissue components from a well established training dataset [3] exhibit relatively subtle differences on an absolute scale. (C) Mean spectra of crypts and tumor regions, with shaded areas around the mean spectra indicating standard deviation. Spectral variability within each class is small even in relation to the subtle differences between average spectra, so that differences between classes (here exemplified by crypts vs. tumor) remain distinguishable by classifiers.