A Face Recognition Siamese Network implemented using Keras. Siamese Network is used for one shot learning which do not require extensive training samples for image recognition.
Using opencv to collect the anchor and positive images and using the lfw dataset to collect the negative images.
scaling and resizing the image to (100,100,3). Creating the labelled dataset i.e. (anchor,positive)--> 1,1,1,1 and (anchor,negative) --> 0,0,0,0
Building embedding layer.
Building the distance layer.
Making the Siamese model.
Setup loss and optimizer
Establishing checkpoints for the training_checkpoints
Build train step funtion and the training loop
Training the model.
Importing the metrics.
Make predictions.
Calculate metrics.
Visualise results.
Saving the model as an h5 file.
real time veirification using webcam.