Facial recognition works by identifying and measuring facial features. This procededure can identify human faces in pictures or live video streams, determine if the face in different images belongs to the same person, or check for a face in a large collection of existing images.
Stonito Faces use a simple procedure for facial recognition. Every face detected is scanned for its unique facial characteristics. This is some kind of face encoding.
This face encoding is compared with other face encodings and the measure of it is in fact a measure of a similarity of the two faces.
At the same time for each detected face software performs the simple check of age and gender and those results are also stored for possible statistical use.
Technology Used
We used well-known and widely used open source libraries:
OpenCV, loading image sources
Dlib, a toolkit for making real world machine learning and data analysis
Those technology are ported to .NET version 6.
The final product works as a standard Windows application for 64 bit.
