A method for analyzing a digital photograph to identify skin areas in a photograph to assist in skin cancer assessment is disclosed. The photograph is transformed to a plurality of colour base candidates from which one is chosen based on the highest score plane. A segmentation of said photograph is then segmented into a plurality of objects based on maximal separations between different objects and these are classified into one or more of the following categories: skin, background or underwear. The colour bases may be: hue, saturation, lightness(HSL), hue, saturation, value(HSV), hue-saturation-intensity(HSI), La*b*(CIELa*b*) or RGB. The segmentation may comprise detecting density thresholds and maximization of interclass variation and the classification may be based on computing the variance for randomly selected label regions.
