Discrimination of inflammatory bowel disease using Raman spectroscopy and linear discriminant analysis methods

Document Type

Conference Proceeding

Publication Date

1-1-2016

Journal

Progress in Biomedical Optics and Imaging - Proceedings of SPIE

Volume

9704

DOI

10.1117/12.2225299

Keywords

Crohn's Colitis; feature selection; Inflammatory Bowel Disease; multivariate analysis; Raman spectroscopy; Ulcerative Colitis

Abstract

© 2016 SPIE. Inflammatory bowel disease (IBD) is an idiopathic disease that is typically characterized by chronic inflammation of the gastrointestinal tract. Recently much effort has been devoted to the development of novel diagnostic tools that can assist physicians for fast, accurate, and automated diagnosis of the disease. Previous research based on Raman spectroscopy has shown promising results in differentiating IBD patients from normal screening cases. In the current study, we examined IBD patients in vivo through a colonoscope-coupled Raman system. Optical diagnosis for IBD discrimination was conducted based on full-range spectra using multivariate statistical methods. Further, we incorporated several feature selection methods in machine learning into the classification model. The diagnostic performance for disease differentiation was significantly improved after feature selection. Our results showed that improved IBD diagnosis can be achieved using Raman spectroscopy in combination with multivariate analysis and feature selection.

This document is currently not available here.

Share

COinS