The present invention provides a method for automatically identifying a person comprises acquiring white-light and UV images of a portion of the person's skin, generating a skin mask from the white-light image, and comparing the skin mask with a pre-stored skin mask of the person. If a substantial match is not found between the two skin masks, the person is not identified, and an error message such as “wrong person” or “person unknown” is returned. Otherwise, the method proceeds to obtain results associated with certain skin conditions using at least the UV image. The results are compared with pre-stored results to determine if the person is the right person or wrong person. A computer readable medium and system are also disclosed.