Challenges of High Melanin Content on Skin Cancer Diagnosis by Hyperspectral Imaging

Abstract

Melanin is an important absorber in skin as well as benign and malignant skin lesions. Due to these effects, the intensity of remitted light decreases and changes the received spectra significantly. Even though melanin itself is highly relevant in distinguishing skin lesions, it is also necessary to find differences beside melanin for distinguishing benign from malignant skin lesions. We investigated the effect that melanin has on two classical algorithms for tissue component extraction using simulated spectra. Our findings demonstrate that high melanin content complicates the extraction of tissue information like oxygenation and component concentration from the imaged areas.