Share Email Print

Journal of Biomedical Optics • Open Access

Determination of uncertainty in parameters extracted from single spectroscopic measurements
Author(s): Obrad R. Scepanovic; Kate L. Bechtel; Abigail S. Haka; Wei-Chuan Shih; Tae-Woong Koo; Andrew J. Berger; Michael S. Feld

Paper Abstract

The ability to quantify uncertainty in information extracted from spectroscopic measurements is important in numerous fields. The traditional approach of repetitive measurements may be impractical or impossible in some measurements scenarios, while chi-squared analysis does not provide insight into the sources of uncertainty. As such, a need exists for analytical expressions for estimating uncertainty and, by extension, minimum detectable concentrations or diagnostic parameters, that can be applied to a single noisy measurement. This work builds on established concepts from estimation theory, such as the Cramér-Rao lower bound on estimator covariance, to present an analytical formula for estimating uncertainty expressed as a simple function of measurement noise, signal strength, and spectral overlap. This formalism can be used to evaluate and improve instrument performance, particularly important for rapid-acquisition biomedical spectroscopy systems. We demonstrate the experimental utility of this expression in assessing concentration uncertainties from spectral measurements of aqueous solutions and diagnostic parameter uncertainties extracted from spectral measurements of human artery tissue. The measured uncertainty, calculated from many independent measurements, is found to be in good agreement with the analytical formula applied to a single spectrum. These results are intended to encourage the widespread use of uncertainty analysis in the biomedical optics community.

Paper Details

Date Published: 1 November 2007
PDF: 10 pages
J. Biomed. Opt. 12(6) 064012 doi: 10.1117/1.2815692
Published in: Journal of Biomedical Optics Volume 12, Issue 6
Show Author Affiliations
Obrad R. Scepanovic, Massachusetts Institute of Technology (United States)
Kate L. Bechtel, Massachusetts Institute of Technology (United States)
Abigail S. Haka, Massachusetts Institute of Technology (United States)
Wei-Chuan Shih, Massachusetts Institute of Technology (United States)
Tae-Woong Koo, Massachusetts Institute of Technology (United States)
Andrew J. Berger, Univ. of Rochester (United States)
Michael S. Feld, Massachusetts Institute of Technology (United States)

© SPIE. Terms of Use
Back to Top