This is the web server written in Java based on OpenCV and Tensorflow Calculations.
This server accepts GET/POST queries with images and returns outcome with identifying prediction based on face features (FaceNet neural network).
Alongside with simple identification it provides finding all registered users on photo and putting labels on them with square on their faces.
Registration is based on jdbc SQL connection. (PostgreSQL, MySQL etc.). Every non-login query requires token authentication based on RSA256.
Server supports updating avatars, returning simple image with all faces marked, face coordinates of all found faces, eyes coorinates etc.
------------------------------EXAMPLE-----------------------------------
First we register new user using API.
POST query
To create a user we must give 3 types of images: where face is looking left, right and straight. We can apply many images, and server will pick mean values of them.
Server created json representation of images (based on 128 face features) that has been passed through Convolutional Neural Network FaceNet and disperses them into 3 categories: left, center and right.
Example of JSON: { left {"features":[0.017909864, ...]}, center{ "features":[-0.014903864,...]} right {"features":[0.4253545443, ....], facelabel: "example_user" }
Then we can send login request, get a token and after than we can send POST query to