Human Face Spoofing Detection

• Github URL: Project Link

Proposed a Convolutional Neural Network (CNN) model to prevent attacks on face recognition systems caused by human face spoofing.

Trained the model using just 15k training samples and attained a test accuracy of 87% on the HKBU-MARs anti-spoofing dataset.

Ensured that the model is lightweight for implementation on a Raspberry Pi device for real-time spoofing detection using a web cam.