Algorithms from Signal and Data Processing Applied to Hyperspectral Analysis: Discriminating Normal and Malignant Microarray Colon Tissue Sections Using a Novel Digital Mirror Device System
Abstract
Hyperspectral imaging is an important tool in various fields, notably geosensing and astronomy, and with the development of new devices, it is now also being applied also in medicine. Concepts and tools from signal processing and data analysis need to be employed to analyze these large and complex data sets. In this paper, we present several techniques which generally apply to hyperspectral data, and we use them to analyze a particular data set. With light sources of increasingly broader ranges, spectral analysis of tissue sections has evolved from 2 wavelength image subtraction techniques to Raman near infra-red microspectroscopic analysis permitting discrimination of cell types and tissue patterns. We have developed and used a unique tuned light source based on micro-optoelectromechanical systems (MOEMS) and applied algorithms for spectral microscopic analysis of normal and malignant colon tissue. We compare the results to our previous studies which used a tunable liquid filter light source.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2004
- Accession Number
- ADA635870
Entities
People
- Andreas C. Coppi
- Frank B. Geshwind
- Frederick J. Warner
- Gustave L. Davis
- Mauro Maggioni
- Richard A. Deverse
- Ronald R. Coifman
Organizations
- Yale University