Face Recognition is a very sophisticated form of Artificial Intelligence and the kind we provide is based on Machine Learning. A necessary step in any Machine Learning process is called “Feature Extraction”. That’s where the computer decides what features it’s going to use to learn from. In the case of a human face, a primary consideration is the distance between one part of the face and another.
This link runs on Tensorflow, a Machine Learning platform from Google. Cloudastructure uses Tensorflow for our Object Detection. However, we use something even more sophisticated for Face Recognition called ArcFace (even more details here
). The “click to see your face
” link is hosted by Google and is a great demonstration of what feature extraction looks like on a live face. We can’t show you exactly what feature extraction looks like on ArcFace because that layer is hidden from direct view, but this more rudimentary platform shows it well. We hope you enjoy the demonstration.