Recently, biometric techniques are used widely in field of security system. Especially, face recognition is
one of popular biometrics due to its intuitiveness and convenience. However, human face is always exposed
and likely to be a target of spoofing attack. Thus, in face recognition security system, fake face detection
is important issue as much as face recognition. In general, human skin color is vary under the influence of
blood flow but it is impossible to perceive such variation with naked eye. Fortunately, a proposed skin
color magnification method makes it possible to capture these variations. From that, we extract the
physiological signal from face skin then distinguish the captured face by camera whether the face image is
live or not. Proposed method has no need to training process before use and does not require user to any
behavior or facial expression. Experiment result showed that the real and fake(facial photo) has distinctive
results in terms of inflection point of the extracted physiological signal.