An algorithm for estimation of background signal of Raman spectra from biological cell samples using polynomial functions of different degrees

Raman microspectroscopy is extensively used to investigate the composition and structure of a wide range of materials at microscopic level. In particular, it is increasingly applied to study the chemical content of biomedical samples for diagnostic purposes. However, so that Raman microspectroscopy is able to provide reliable results, the removal of spectral background signals, due to both intrinsic autofluorescence and stray light, which are usually a few orders of magnitude larger than those related to Raman signal, should be correctly achieved. A semiautomated method to remove broad background from Raman spectra of cellular samples is described. It requires a slight user intervention for controlling that several spectral parameters are located away from Raman peaks. These discrete data are used to separate the spectrum in different spectral regions and a proper polynomial function approximates the background for each spectral region. The method yields acceptable and interesting results when applied both to synthetic and experimental Raman spectra.