Correction for Extrinsic Background in Raman Hyperspectral Images

Raman hyperspectral microscopy is a valuable tool inbiologicaland biomedical imaging. Because Raman scattering is often weak incomparison to other phenomena, prevalent spectral fluctuations andcontaminations have brought advancements in analytical and chemometricmethods for Raman spectra. These chemometric advances have been keycontributors to the applicability of Raman imaging to biological systems.As studies increase in scale, spectral contamination from extrinsicbackground, intensity from sources such as the optical componentsthat are extrinsic to the sample of interest, has become an emergingissue. Although existing baseline correction schemes often reduceintrinsic background such as autofluorescence originating from thesample of interest, extrinsic background is not explicitly considered,and these methods often fail to reduce its effects. Here, we showthat extrinsic background can significantly affect a classificationmodel using Raman images, yielding misleadingly high accuracies inthe distinction of benign and malignant samples of follicular thyroidcell lines. To mitigate its effects, we develop extrinsic backgroundcorrection (EBC) and demonstrate its use in combination with existingmethods on Raman hyperspectral images. EBC isolates regions containingthe smallest amounts of sample materials that retain extrinsic contributionsthat are specific to the device or environment. We perform classificationboth with and without the use of EBC, and we find that EBC retainsbiological characteristics in the spectra while significantly reducingextrinsic background. As the methodology used in EBC is not specificto Raman spectra, correction of extrinsic effects in other types ofhyperspectral and grayscale images is also possible.

Automated summary

Raman hyperspectral microscopy is a valuable tool in biological and biomedical imaging. However, contamination from extrinsic background significantly affects classification models. To mitigate this issue, researchers developed an extrinsic background correction method and demonstrated its use in combination with existing methods on Raman hyperspectral images.