Face spoofing indicator using deep learning

Facial recognition is a popular biometric authentication method because of its convenience and lack of physical interaction by the end-user. However, facial recognition systems are vulnerable to face spoof attacks because of the ease to acquire people’s photos from social networking sites. Therefore, this project aims to tackle 2D face spoofing attacks by developing and training a deep learning model that can differentiate real and spoofed faces. The liveness detection model was trained with the collected image dataset so that it could classify and predict face detections into 2 classes, real and fake. The results showed that the model had an accuracy close to 100% that could differentiate real and spoofed faces from the video stream of the laptop’s web camera.Bachelor of Engineering (Information Engineering and Media