Noise in biological Raman spectroscopy

Raman spectroscopy is a widely used method to investigate chemical molecules by analyzing their vibrational transitions. It utilizes inelastic scattering of the laser light irradiating the investigated object. The scattered light requires appropriate filtering to reduce dominant laser light and expose much weaker components having shifted wavelengths of a characteristic spectral pattern. These components are measured by dispersing onto a detector. The Raman scattered light is very weak and limits applications of Raman spectroscopy. The technique can be potentially applied to investigate in vivo biological tissues to provide real-time diagnosis of serious diseases. One of limitations of the method is caused by problems of identification of a very weak scattered light component at the presence of inherent noise of the measurement setup and external interferences. Usually, the main source of noise in a system is a detector. Various methods can be proposed to reduce its inherent noise. The Raman spectra measurement systems were presented in detail to identify noise sources. The main sources of the detector inherent noise were presented. Next, the signal processing methods, reducing the additive noise component in Raman spectra, were described. Finally, an influence of additive noise in Raman spectra on selected detection algorithms was considered. The results determine how the additive noise component influences concentration accuracy of tissue constituents, identified by popular regression algorithms.