Face recognition using FaceNet, and for security we add eye blinking detection for detecting fake faces.
This project is composed to 3 parts:
FaceNet learns a neural network that encodes a face image into a vector of 128 numbers.
By comparing two such vectors, you can then determine if two pictures are of the same person.
Encoding model : Inception
Loss: Triplet loss
Face recognition systems can be circumvented simply by holding up a photo of a person to the face recognition camera.
In order to make face recognition systems more secure, we need to be able to detect such fake/non-real faces using Liveliness Net.
Problem: Binary classification of eye status.
Dataset : Closed Eyes In The Wild (CEW) dataset
Model : LeNet-5
Combine the two previous networks, and implementing real time feature using opencv.
Idea : We won’t confirm the face if it’s detected until the eyes blink.
Face detection task : openCV pre-trained Haar-cascade classifier.