Selecting optimal features from Fourier transform infrared spectroscopy for discrete-frequency imaging

Rupali Mankar, Michael J. Walsh, Rohit Bhargava, Saurabh Prasad, David Mayerich

Research output: Contribution to journalArticle

Abstract

Tissue histology utilizing chemical and immunohistochemical labels plays an important role in biomedicine and disease diagnosis. Recent research suggests that mid-infrared (IR) spectroscopic imaging may augment histology by providing quantitative molecular information. One of the major barriers to this approach is long acquisition time using Fourier-transform infrared (FTIR) spectroscopy. Recent advances in discrete frequency sources, particularly quantum cascade lasers (QCLs), may mitigate this problem by allowing selective sampling of the absorption spectrum. However, DFIR imaging only provides a significant advantage when the number of spectral samples is minimized, requiring a priori knowledge of important spectral features. In this paper, we demonstrate the use of a GPU-based genetic algorithm (GA) using linear discriminant analysis (LDA) for DFIR feature selection. Our proposed method relies on pre-acquired broadband FTIR images for feature selection. Based on user-selected criteria for classification accuracy, our algorithm provides a minimal set of features that can be used with DFIR in a time-frame more practical for clinical diagnosis.

Original languageEnglish (US)
Pages (from-to)1147-1156
Number of pages10
JournalAnalyst
Volume143
Issue number5
DOIs
StatePublished - Mar 7 2018

Fingerprint

Histology
histology
Fourier Transform Infrared Spectroscopy
FTIR spectroscopy
Fourier transform infrared spectroscopy
Feature extraction
Infrared radiation
Imaging techniques
Semiconductor Lasers
Quantum cascade lasers
Discriminant Analysis
Discriminant analysis
Fourier Analysis
absorption spectrum
discriminant analysis
genetic algorithm
Fourier transform
Labels
Absorption spectra
Fourier transforms

ASJC Scopus subject areas

  • Analytical Chemistry
  • Biochemistry
  • Environmental Chemistry
  • Spectroscopy
  • Electrochemistry

Cite this

Selecting optimal features from Fourier transform infrared spectroscopy for discrete-frequency imaging. / Mankar, Rupali; Walsh, Michael J.; Bhargava, Rohit; Prasad, Saurabh; Mayerich, David.

In: Analyst, Vol. 143, No. 5, 07.03.2018, p. 1147-1156.

Research output: Contribution to journalArticle

Mankar, Rupali ; Walsh, Michael J. ; Bhargava, Rohit ; Prasad, Saurabh ; Mayerich, David. / Selecting optimal features from Fourier transform infrared spectroscopy for discrete-frequency imaging. In: Analyst. 2018 ; Vol. 143, No. 5. pp. 1147-1156.
@article{d4069f9d6cc64fedbaa1ba3a60b7dd52,
title = "Selecting optimal features from Fourier transform infrared spectroscopy for discrete-frequency imaging",
abstract = "Tissue histology utilizing chemical and immunohistochemical labels plays an important role in biomedicine and disease diagnosis. Recent research suggests that mid-infrared (IR) spectroscopic imaging may augment histology by providing quantitative molecular information. One of the major barriers to this approach is long acquisition time using Fourier-transform infrared (FTIR) spectroscopy. Recent advances in discrete frequency sources, particularly quantum cascade lasers (QCLs), may mitigate this problem by allowing selective sampling of the absorption spectrum. However, DFIR imaging only provides a significant advantage when the number of spectral samples is minimized, requiring a priori knowledge of important spectral features. In this paper, we demonstrate the use of a GPU-based genetic algorithm (GA) using linear discriminant analysis (LDA) for DFIR feature selection. Our proposed method relies on pre-acquired broadband FTIR images for feature selection. Based on user-selected criteria for classification accuracy, our algorithm provides a minimal set of features that can be used with DFIR in a time-frame more practical for clinical diagnosis.",
author = "Rupali Mankar and Walsh, {Michael J.} and Rohit Bhargava and Saurabh Prasad and David Mayerich",
year = "2018",
month = "3",
day = "7",
doi = "10.1039/c7an01888f",
language = "English (US)",
volume = "143",
pages = "1147--1156",
journal = "Analyst",
issn = "0003-2654",
publisher = "Royal Society of Chemistry",
number = "5",

}

TY - JOUR

T1 - Selecting optimal features from Fourier transform infrared spectroscopy for discrete-frequency imaging

AU - Mankar, Rupali

AU - Walsh, Michael J.

AU - Bhargava, Rohit

AU - Prasad, Saurabh

AU - Mayerich, David

PY - 2018/3/7

Y1 - 2018/3/7

N2 - Tissue histology utilizing chemical and immunohistochemical labels plays an important role in biomedicine and disease diagnosis. Recent research suggests that mid-infrared (IR) spectroscopic imaging may augment histology by providing quantitative molecular information. One of the major barriers to this approach is long acquisition time using Fourier-transform infrared (FTIR) spectroscopy. Recent advances in discrete frequency sources, particularly quantum cascade lasers (QCLs), may mitigate this problem by allowing selective sampling of the absorption spectrum. However, DFIR imaging only provides a significant advantage when the number of spectral samples is minimized, requiring a priori knowledge of important spectral features. In this paper, we demonstrate the use of a GPU-based genetic algorithm (GA) using linear discriminant analysis (LDA) for DFIR feature selection. Our proposed method relies on pre-acquired broadband FTIR images for feature selection. Based on user-selected criteria for classification accuracy, our algorithm provides a minimal set of features that can be used with DFIR in a time-frame more practical for clinical diagnosis.

AB - Tissue histology utilizing chemical and immunohistochemical labels plays an important role in biomedicine and disease diagnosis. Recent research suggests that mid-infrared (IR) spectroscopic imaging may augment histology by providing quantitative molecular information. One of the major barriers to this approach is long acquisition time using Fourier-transform infrared (FTIR) spectroscopy. Recent advances in discrete frequency sources, particularly quantum cascade lasers (QCLs), may mitigate this problem by allowing selective sampling of the absorption spectrum. However, DFIR imaging only provides a significant advantage when the number of spectral samples is minimized, requiring a priori knowledge of important spectral features. In this paper, we demonstrate the use of a GPU-based genetic algorithm (GA) using linear discriminant analysis (LDA) for DFIR feature selection. Our proposed method relies on pre-acquired broadband FTIR images for feature selection. Based on user-selected criteria for classification accuracy, our algorithm provides a minimal set of features that can be used with DFIR in a time-frame more practical for clinical diagnosis.

UR - http://www.scopus.com/inward/record.url?scp=85042693851&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042693851&partnerID=8YFLogxK

U2 - 10.1039/c7an01888f

DO - 10.1039/c7an01888f

M3 - Article

C2 - 29404544

AN - SCOPUS:85042693851

VL - 143

SP - 1147

EP - 1156

JO - Analyst

JF - Analyst

SN - 0003-2654

IS - 5

ER -