vadikko2/ssp

Face detection and recognition with a Telegram bot as a control panel

Requirements:

• python 3.* and these specific packages

    $ pip3 install python-telegram-bot
    $ pip3 install Pillow OR sudo apt install python3-pil
  

• OpenCV 3.*

How to run:

Create training sample of a person

$ mkdir db && cd ./db && mkdir <name>

$ mkdir faces
$ python3 data.py <name>

To train a chosen recognizer (Eigen, Fisher or LBPH)

$ python3 train.py recognizer

Run a server with a chosen recognizer (Eigen, Fisher or LBPH) and Telegram bot packed in. Don't forget to insert your bot token first

$ python3 main.py recognizer port_number

Run a client with an ultrasonic sensor connected (Intel Edison is preferred)

$ python3 ./client/detect.py port_number

Bot commands:

Telegram bot has two commands:

• /stay
• /stop

/stay

Adds a job which is going to be executed every time detection occurs. If so happened, face_recognition() method is going to try to find any faces and predict who is on a picture, then it will send it to your Telegram chat with a bot.

/stop

Removes previously added job