Camigo is a basic python app that attempts to recognize people faces streamed via a camera (or webcam) using haar cascades.
Pre-requisites: python 2.7, opencv, numpy
Usage: python camigo.py
When application runs, it presents 4 menu options:
- take photo
- train system
- arm system
- exit
To set up someone's face, select option 1. This will prompt to enter the person's name (specify with no spaces). The system will take a set up pictures (50) and store them in a folder named after the person. Repeat this procedure to setup a different person.
Note that the higher the number of pictures and different facial expressions, the better the accuracy.
Once everyone's pictures are taken, select option 2 to train the system.
Then select option 3 to arm the system which will start the camera and display a green rectangle around people's faces with their name. If the system cannot recognise the person, it will display "intruder" as the name.
One of the use cases of interest is to set it up as a surveillance camera and adapt it to send sms, emails, notification (with attached images) based on certain events such as intruder detection (unrecognised faces), motion detection etc. (with sensors) using a raspberry pi.