Face recognition is able to identify or verify the identity of the main body of technology of the image or video.First human face recognition algorithm, was born in the early seventy s.Since then, their accuracy has been improved, compared to now such as fingerprints or iris recognition traditionally considered more stable biological recognition method, people tend to prefer face recognition.Let facial recognition is more popular than other biometric methods one big difference is face recognition are invasive nature.Fingerprint identification, for example, users need to finger on the sensor, the iris recognition need a camera and the user in close proximity, speech recognition requires users to speak loudly.Relatively modern face recognition system only requires the user to within the camera view (assuming they also reasonable and the distance of the camera).This makes face recognition has become the most friendly to the user biological recognition method.This also means that the potential application of face recognition is wider, because it can also be deployed in the user does not expect to cooperate with system environment, such as monitoring system.Face recognition of other common applications include access control, fraud detection, identification and social media.
when deployed in the environment without constraints, because of the face image presented has a high degree of variability in the real world, this kind of face image is often referred to as natural face (faces in - the - wild)), so face recognition is one of the most challenging biological recognition method.Variable places including head posture, facial image age, shade, light conditions, and facial expressions.Figure 1 shows the situation of the sample.