Iot Face Recognition System

This branch is the basic version of the system which can run in the PC. The web server version is in the other branch. The front end of this project is here.

Function

Face Detection

• Store Your Faces
  When the program runs, people can press

  • key s to store and restore their faces
  • key r to remove the faces with specific name

• Faces Detection
  The program can detect faces in the camera and use square the around the faces. Multi-faces are supported.

• Faces Recognition
  The camera can recognize who are in the camera. Multi-faces are supported.The labels will be given near the square. The labels are

  • Unknown if the faces hasn't been stored
  • the name that people gave when they store the faces

IoT Control

• Send Message to mobiles
  When strangers come to the house, a message via Facebook Messager is given to their mobiles.

• Unlock Control
  When the house owner get back home, an unlock key image will be shown in the screen.

Install

Software Requirments

• Install face_recognition package
  To install this package in Windows

  • if you already have the cmake compiler environment, you can use following command to install the package

    pip install dlib
    pip install face_recognition
    

  • if you don't have the camke compiler environment, you can go to here to download the dlb.whl file and install it with pip. Later you can use

    pip install face_recognition
    

    to install face_recognition package.

  I personal recommend you can use the virtualenv to create a virtual python environment, instruction is here.

• IFTTT settings

  • Signup an account in IFTTT website.
  • Go to MyApplets - New Applet - this, choose the Webhooks - Revice a web request, and give the event name and click Create trigger.
  • choose that - Face Messenger - Send Message, and type what you want to send to your mobile, then click Create Action - Finish.
  • Go back to the homepage and go to MyApplets - Services - Webhooks
  • Documentation copy the link to the parameter of url_stranger in home_assistance_robot.py

• Use PyCharm to open this project, make sure that your interpreter was set to the right path of your path and dlib, face_recogniton were already installed.

• Run home_assistance_robot.py

Hardware Requirements

• A PC with python environment
• A Camera connected to the PC
• A Mobile with Facebook Messenger

Structure and Parameters

• main.py
  The entry of all the functions. Use the FaceProcess class from the ImageProcess.py file.

• ImageProcess.py
  Has the class of FaceProcess, can handle the function of camera detection and recognition.
  You can set the parameters, the parameters are as following:

  • resize_frame
    Resize the frame when handle the frame. Vary from 0 to 1.
  • recognize_threshold
    The threshold to recognize people. Vary from 0 to 1.
  • recognize_mode
    0 means to recognize when the number of faces change; 1 means to recognize every detect_interval frames.
  • detect_interval
    handle the image every detect_interval frames. Has no function if recognize_mode is set to 0
  • person_store_number
    the number of images to store in the database
  • filename=filename
    the .csv file name

  Notes:
  I recommend you set the parameter to recognize_mode=1, detect_interval=2. If it runs very slowly, you can set the parameter to recognize_mode=0.