In this paper we propose a new automatic identification system. In our system, a user?s face is captured by a camera in order to identify him/her. By using an illumination difference image which is made of light coming from the right and left sides, our system can distinguish between a real person and a photograph. We describe an algorithm that detects whether the object facing the camera is planar (in the case of a photograph) or not (in the case of a real person). We propose the technique of performing person-photograph distinction by the surface direction image features.
Citation:
Takahiro Tsuda, Kazuhiko Yamamoto, Kunihito Kato, "Consideration of a Person/Photograph Distinction Using the Surface Direction Image," crv, pp.195-200, 1st Canadian Conference on Computer and Robot Vision (CRV'04), 2004