Using multiple wave-bands is common in multispectral imaging and remote sensing, e.g., to enable improved reproduction of colours or to detect the crop status in a field, respectively. Computer vision methods, however, mainly rely on monochrome or RGB wave-bands, although segmentation tasks might be improved considerably using more or other wave-bands than the standard RGB.This paper presents and investigates a new approach for detecting human skin using a combination of standard RGB bands and three near infrared bands. Simulations under changing illumination conditions and a preliminary experiment show an improved robustness over pure RGB based approaches.
Moritz Störring, Hans J. Andersen, Erik Granum, "A multispectral approach to robust human skin detection" in Proc. IS&T CGIV 2004 Second European Conf. on Colour in Graphics, Imaging, and Vision, 2004, pp 110 - 115, https://doi.org/10.2352/CGIV.2004.2.1.art00024