On a taxonomy of facial features

After nearly a decade of intensive research in face recognition, no standard organization exists for grouping the salient information available in 2D face images into feature categories. At the same time, human verification of a subject's identity based on facial images lacks a consistent methodology. In this paper we propose a taxonomy of available facial features that: (i) serves as a precursor to studies on the individuality of facial features, (ii) follows a similar well established and accepted organization for fingerprint features, and (iii) contains features computable by both machines and humans as well as by machines alone. This manuscript is intended as a strawman of an organization of facial features, that would hopefully lead to a standardization of such features. Such a facial feature organization will (i) enable studies on the individuality of facial features, which has important ramifications for the acceptance of expert testimony in legal proceedings for determining the identity of an individual from a facial photograph, and (ii) help standardize the framework of commercial face recognition systems.