How to activate
the Face Recognition function using the Robot Benri.
IMPORTANT
1- This system works like a human being, the more times you see
one person, the better the recognition of that person.
So that, the more faces you register of one person, the best the
recognition results.
2- The face recognition system needs to be trained in the
following way: First register one face of one
user. In case of obtaining a bad recognition of that
person, you need to register one more face of that person.
Each time a bad recognition of one person is obtained, you need
to register one more face of that person. Do not
register new faces of one person, when correct recognition
results are obtained.
3- When one person registers his face, he needs to look directly
to the camera. Glasses needs to be avoided and any
other element that can occlude one person's eyes.
4- This method works by recognizing first one person's eyes.
The method looks for both eyes of one person. If one
eye is occluded, this system will not be able to recognize that
person.
5- The number of faces to be registered, depends on the level of
the light variation around your cameras.
6- Environments having a low level of light variation, i.e.
located far away from windows or any other sources that can
produce a high level of light variation, will need a less amount
of faces to be registered. This environments are very much
recommended for your cameras.
7- Environments having a high level of light variation, i.e.
located near windows or any other sources that can produce a
high level of light variation, will need a more amount of faces
to be registered. This environments are not
recommended for your cameras.
8- The larger the face size, the best the recognition results.
9- The best the camera quality, the best the recognition
results.
10- The best the illumination around your cameras, the best the
recognition results.
