Performance characterization of material identification systems

Christopher D. Brown; Robert L. Green

doi:10.1117/12.686103

25 October 2006 Performance characterization of material identification systems

Christopher D. Brown, Robert L. Green

Proceedings Volume 6378, Chemical and Biological Sensors for Industrial and Environmental Monitoring II; 637809 (2006) https://doi.org/10.1117/12.686103
Event: Optics East 2006, 2006, Boston, Massachusetts, United States

Abstract

In recent years a number of analytical devices have been proposed and marketed specifically to enable field-based material identification. Technologies reliant on mass, near- and mid-infrared, and Raman spectroscopies are available today, and other platforms are imminent. These systems tend to perform material recognition based on an on-board library of material signatures. While figures of merit for traditional quantitative analytical sensors are broadly established (e.g., SNR, selectivity, sensitivity, limit of detection/decision), measures of performance for material identification systems have not been systematically discussed. In this paper we present an approach to performance characterization similar in spirit to ROC curves, but including elements of precision-recall curves and specialized for the intended-use of material identification systems. Important experimental considerations are discussed, including study design, sources of bias, uncertainty estimation, and cross-validation and the approach as a whole is illustrated using a commercially available handheld Raman material identification system.

Citation Download Citation

Christopher D. Brown and Robert L. Green "Performance characterization of material identification systems", Proc. SPIE 6378, Chemical and Biological Sensors for Industrial and Environmental Monitoring II, 637809 (25 October 2006); https://doi.org/10.1117/12.686103

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available