Artificial intelligence is the future of modern age
I revisit both the alignment step and the representation step by employing explicit 3D face modeling.
DeepFace shows a very close performance to the human level. Humans have 97.53% score whereas the DeepFace model has 97.35% ± 0.25%. This means that the model can get a higher score than human beings sometimes. I am providing the DeepFace model within Keras for Python in this gig.
This system is a combination of CNN, auto-encoder, and transfer learning.
I trained the model on the largest facial dataset to-date (SFC Dataset). The data set has 4.4 million of photos of 4030 people. That is exactly the number of nodes in the final layer – F8
I tested the developed model with variations of Angelina Jolie and Jennifer Aniston. Surprisingly, the model can verify all instances I fed. You can use this model in real-time as well