A robust method for automated background subtraction of tissue fluorescence

✍ Scribed by Alex Cao; Abhilash K. Pandya; Gulay K. Serhatkulu; Rachel E. Weber; Houbei Dai; Jagdish S. Thakur; Vaman M. Naik; Ratna Naik; Gregory W. Auner; Raja Rabah; D. Carl Freeman

• Publisher: John Wiley and Sons
• Year: 2007
• Tongue: English
• Weight: 482 KB
• Volume: 38
• Category: Article
• ISSN: 0377-0486
• DOI: 10.1002/jrs.1753

✦ Synopsis

Abstract

This paper introduces a new robust method for the removal of background tissue fluorescence from Raman spectra. Raman spectra consist of noise, fluorescence and Raman scattering. In order to extract the Raman scattering, both noise and background fluorescence must be removed, ideally without human intervention and preserving the original data. We describe the rationale behind our robust background subtraction method, determine the parameters of the method and validate it using a Raman phantom against other methods currently used. We also statistically compare the methods using the residual mean square (RMS) with a fluorescence‐to‐signal (F/S) ratio ranging from 0.1 to 1000. The method, ‘adaptive minmax’, chooses the subtraction method based on the F/S ratio. It uses multiple fits of different orders to maximize each polynomial fit. The results show that the adaptive minmax method was significantly better than any single polynomial fit across all F/S ratios. This method can be implemented as part of a modular automated real‐time diagnostic in vivo Raman system. Copyright © 2007 John Wiley & Sons, Ltd.