Learning how to use face_recognition and dlib python libraries.
Software menu:
1. Listar Pessoas
2. Registrar
3. Start
4. Deletar
5. Salvar e Sair
0. Dev Tools
DEV TOOLS
1. Toggle Debugging
2. Set Confianca
3. Resetar log
4. Resetar todos os arquivos
5. Voltar para menu
This script is capable of recognizing, at real time, people's faces and who they belong to. It is also possible to register new people pressing SPACE - while on program 2. - however many times you wish to train a picture (the more pictures, the more accurate the system will be, but also slower), then pressing ESC and naming the person. You can also delete previously registered people and change the threshold to associate a face with someone (default is .4)
Whenever unknown faces are present in the frame, a counter will increase and, upon reaching an user defined value (default is ~6 seconds), will take a snapshot of the frame every ~6 seconds and save it to safe/{datetime}.png
The system also mantains a log file, which registers new and deleted users, as well as when snapshots were taken and when some known face entered or exited the frame
face_recognition (requires dlib)
Don't know if it works on windows! (not officially supported by face_recognition)
Just download the .py file and run it in an environment with the required libraries and a folder named "safe" at the same location as the script, to store snapshots.
Three files will be created, if non-existant, named nomes.txt - which stores every registered name in binary form - log.txt - which can be opened with a text editor to be read - and encodings.npy - also stored in binary, this is the array with every previously trained face.
Remember to always exit the program pressing 5. and ENTER, otherwise it will not save any file changes
Ubuntu 18.04 |
|
face_recognition 1.2.2 |
|
dlib 19.15.0 |
works! |
openCV 3.4.2 |
|
Python 3.6.5 |
[ ] Add possibility to train on new faces from local image set