LIVENESS DETECTION WITH OPENCV

Biometric systems are commonly used to identify authorised individuals based on behavioural or physical features. However, various traits can be used to fake this. Spoofing is defined as the use of security features to attack or harm a biometric recognition system in order to use it without the authorization of an authorised user. Images or video of a person can readily be found on social media or shot from a safe distance. Consider what would happen if a malicious person tried to evade your face recognition system on purpose. A user like this could try to hold up a photo of someone else. Perhaps they have a snapshot or video on their smartphone that they might show to the camera that performs face recognition. We used OpenCV to detect “fake” versus “real/legitimate” faces in the project and to see how we could use anti-face spoofing techniques in our facial recognition applications by using liveness detection.