DermaSensor's machine-learning algorithm is derived from thousands of spectral samples of pathologically-verified lesions and is trained by expert data scientists. 

Many studies and publications show that malignant tissue, including skin tissue, exhibits distinct spectral signatures with elastic scattering spectroscopy. 

The machine learning aspect of the algorithm is that the algorithm has not been trained to identify pre-determined spectral features, rather the algorithm has "learned" on its own what spectral features and what wavelengths of light to analyze in order to most effectively differentiate malignant from benign skin lesions.