In Door Face Mask Inspector (IDFMI)

Description

1. Overview

This is a project that aims to allow the return of working environment in indoor offices.

It focuses on reading live streams from ip cameras in enterprise environment, after that, it applies a face mask detection model using the state-of-art Artificial Neural Networks.

In addition to detection, it also provides a customization platform for surveillence cameras.

Another service provided from the system is alerting security members via email.

2. Project Architecture

The project consists of 3 containers

idfmi: Contains the Flask application and gunicorn application server
nginx: Contains the Nginx web server
mongo: Contains the MongoDB database

Some info about the stack used

Python version: 3.8.1
Web framework:  Flask
Database:       MongoDB (Using the pymongo driver)
App server:     gunicorn
Worker class:   meinheld
Web server:     Nginx
Front end:      Bootsrap4, JQuery

The system Architecture of IDFMI

3. Project Tree

The following is how the files are structured inside the project:

In-Door-Face-Mask-Inspector/
    .env
    .gitignore
    docker-compose.yml
    LICENSE
    README.md
    app/
        app.py
        camera.py
        config.py
        Dockerfile
        requirements.txt 
        model/
            face_mask_detection.pb
            load_model.py
            utils.py
        static/
            favicon.ico
            main.css     
            main.js   
            main.scss  
        templates/
            base.html
            auth_error.html 
            forms/
               camera.html
               security.html 
            pages/
               home.html
               welcome.html
    mongo/
        Dockerfile
        settings-insert.js
        setup.sh
    nginx/
        Dockerfile
        idfmi.conf
        nginx.conf

4. Screenshots

The home page before signing

• light theme

• dark theme

The home page after signing

• light theme

• dark theme

Different functionalities

• Adding a camera

  • light theme

  

  • dark theme

  

• Adding a security member

  • light theme

  

  • dark theme

  

• View all security member

  • light theme

  

  • dark theme

  

• Update forms are similar to create forms.

Camera Stream and Detection

• light theme

• dark theme

Requirements

1. Docker

Since the project is containerized, then the only important dependency is to have docker and docker compose installed, installation guides are as follows:

• Docker Desktop for Windows: https://docs.docker.com/docker-for-windows/install/
• Docker Desktop for Mac OS: https://docs.docker.com/docker-for-mac/install/
• Docker Engine for linux: https://docs.docker.com/engine/install/
• Docker-Compose for linux: https://docs.docker.com/compose/install/

Notice that for Windows and Mac you should only install docker desktop, which will download docker engine and docker-compose by default.

2. Microsoft Azure Subscription

This is a requirement for the demo only to simulate the working environment with respect to authentication, authorization, and Secrets.

You can check microsoft's documentation on how to make a project with azure active directory and keyvault.

• Azure Active Directory: https://azure.microsoft.com/en-us/services/active-directory/#documentation

• Azure KeyVault: https://azure.microsoft.com/en-us/services/key-vault/#documentation

3. Email Account

• could use any email (gmail, yahoo, outlook), but we used gmail for our prototype.

Installation

• clone the repository:

  git clone https://github.com/othmanKisha/In-Door-Face-Mask-Inspector.git

• After finishing all the steps for Azure Active Directory and Key Vault, make sure to create a .env file inside the In-Door-Face-Mask-Inspector folder and inside the file add the following:

  KEY_VAULT_NAME=${KEY_VAULT_NAME} # Stored in OS environment
  AZURE_CLIENT_ID=${AZURE_CLIENT_ID} # Stored in OS environment
  AZURE_CLIENT_SECRET=${AZURE_CLIENT_SECRET} # Stored in OS environment
  AZURE_SUBSCRIPTION_ID=${AZURE_SUBSCRIPTION_ID} # Stored in OS environment
  AZURE_TENANT_ID=${AZURE_TENANT_ID} # Stored in OS environment
  
  MONGODB_DATABASE=<The name of the database to be created>
  MONGODB_USERNAME=<The username of the database to be created>
  MONGODB_PASSWORD=<The password of the database to be created>
  MONGODB_HOSTNAME=mongodb # This is the name of mongodb container
  MONGODB_PORT=<The port where mongodb will be running> # By default it is 27017
  
  EMAIL_ADDRESS=<The email address created to send alerts>
  EMAIL_PASSWORD=<The password of the alerts email address>
  SMTP_SERVER=<The mail server, it is smtp.gmail.com when you use gmail, you can search the other mail servers>
  SMTP_PORT=<The mail server port> # The port where the smtp server is running by default 587
  
  MODEL_CONFIDENCE=<The confidence of the model> # By default 0.5
  MODEL_FILE=face_mask_detection.pb # The name of the model, can be changed
  MODEL_DIRECTORY=model # The folder containing the model, can be changed

Usage

• make sure you are in the project directory:

  cd In-Door-Face-Mask-Inspector

• run docker-compose command:

  docker-compose up -d

• now you can view the website by typing localhost on the browser.

Acknowledgment

• Face Mask Detection Model by AIZOOTech

  Link: https://github.com/AIZOOTech/FaceMaskDetection

• Camera Streaming with Flask by Miguel Grinberg

  Link: https://github.com/miguelgrinberg/flask-video-streaming

• Azure Active Directory with Flask by René Bremer

  Link:https://github.com/rebremer/ms-identity-python-webapp-backend

Authors

• Abdullah Alnasser

• Mahmoud Ellouh

• Othman Kisha

License

MIT