Unsupervised and supervised methodologies for identification of sample pixels in Fourier transform infrared microspectroscopic images

Abstract

Mid-infrared spectroscopy is a promising label-free technique that can offer insights into morphological and pathological alterations in biological tissues at the molecular level. Owing to the development of the Fourier Transform InfraRed (FTIR) spectrometer, combined with scanning microspectroscopy, FTIR microspectroscopic images can be acquired by measuring spectral data from multiple spatial points, generating comprehensive chemical maps. In data pre-processing, the identification of sample pixels, excluding the background pixels, is important for further effective feature extraction in FTIR images. Herein, we present and compare three methodologies realized using unsupervised and supervised approaches for the identification of the sample pixels and validate them in our experimentally tested FTIR images on tissue sections from multiple organs. The algorithms in the methodologies demonstrate accurate prediction results of the sample and background pixels, and the supervised method further enables automatic detection. These findings highlight thorough and robust solutions to the sample pixel detection problem in FTIR images, contributing to the FTIR signal processing and future chemical and clinical applications of FTIR images.