Multispectral Polarization Imaging for Observing Blood Oxygen Saturation in Skin Tissue

Abstract

We propose a new technique that combines two-dimensional (2D) multispectral imaging and polarization gating for observing the blood oxygen saturation (SpO2) level in human skin tissue. The spectral decomposition of the skin tissue image provides the principal information on blood oxygenation. The polarization gating selects the measurement depth according to the relative orientation of the two polarizers that are placed on a camera and a light source. The combination of these two methods yields multispectral images of the superficial and deep layers of the skin tissue separately. In order to evaluate the blood oxygen, we focus on the multispectral images of the deep site. The SpO2 levels at each image pixel are calculated by means of the partial least squares regression with respect to each reflectance spectrum. The reassignment of the predicted responses retrieves an image whose pixel values represent the relative SpO2 levels. A demonstration experiment for acquiring the multispectral polarization images is performed in the spectral range of 500 to 680 nm, and the SpO2 distributions are obtained.