Iterative morphological and mollifier-based baseline correction for Raman spectra

In vivo Raman spectroscopy with low signal-to-noise ratio and strong, irregularly shaped fluorescence background imposes a challenge for automatic baseline correction methods. In this work, an approach that enables fast and efficient batch baseline correction has been developed, which is based on a morphological operation in combination with a mollifier algorithm. As this algorithm relies only on three parameters, which are determined by the given experimental conditions, it can be used for automatic and objective processing of many Raman spectra. The applicability of the baseline correction is demonstrated on resonance Raman spectra of beta-carotene mixed with fluorescent red ink as model system, on carotenoids in human skin, and on an excitation-emission map of the green alga Haematococcus pluvialis. In the future, the algorithm opens the potential for wide application in Raman spectra analysis in biological contexts. In particular, it greatly facilitates data processing in cases where special photochemical sample preparation or complex experimental baseline removal was required before. Similarly, processing data of experiments using resonant excitation techniques yielding strong fluorescence background is possible. Copyright (c) 2016 John Wiley & Sons, Ltd.