Review of techniques to analyze spectra obtained across diffuse reflectance for diagnosis of skin lesions

Determination of tissue optical properties across of diffuse reflectance spectroscopy is an effective and extensively used technique for the non-invasive study of tissues. The methods to obtain the spectra are practically uniform, however, approaches for signal detection and data processing have varied widely. In some cases it is used a physical model (Monte Carlo model, the diffusion approximation to the transport equation), in other cases an empiric model (principal component analysis) accompanied by tools (Support Vector Machine, Neural Network, Partial Least Squares, etc.) that allow to analyze the obtained spectra. In this work, we present a critical review of the available methods used to analyze the measured diffuse reflection spectra of human skin for diagnostic purposes.