Facial Recognition Project

This project is aimed at developing a air gapped facial recognition system using Python and other various libraries.

It can be found at: Github for Air Gapped Facial Recognition
A downloadable report for this project can be found here: Report

Objectives:

  1. Implementing a facial detection algorithm.
  2. Training a model to recognize faces.
  3. Integrating the trained model into a real-time application.
  4. Deploying the application to various platforms.
  5. Do all this while ensuring the log in module is secure by keeping it air gapped

Project Structure:

The project is organized as follows:

  • src/: Contains the source code for the facial recognition system.
    • facial_detection.py: Script for detecting faces in images or video streams.
    • facial_recognition.py: Script for training and recognizing faces.
    • app.py: Main application script for real-time face recognition.
  • data/: Contains datasets for training the facial recognition model.
    • faces/: Directory containing labeled face images for training.
  • models/: Saved models after training.
  • docs/: Documentation files.
  • requirements.txt: Dependencies required to run the project.
  • etcprof.sh: Login script modified to bypass the main Linux login screen and default to the Raspberry Pi.

Explanation of etcprof.sh snippet:

The provided snippet is a part of the etcprof.sh script, and it performs the following
actions:

if [ "`id -u -n`" == "anvay" ]; then
    ssh -tt tacwin@rpi4.local 2> /dev/null | grep -i "anvay" > /dev/null 
    if [ "$?" -ne "0" ]; then
        now=`date +"%D   %T"`
        echo "login failed at $now" >> /var/attempt_log.log
        exit 1
    else
        now=`date +"%D   %T"`
        echo "login successful at $now" >> /var/attempt_log.log
    fi
fi

  • The if statement checks if the current user executing the script is “anvay”. It uses the id -u -n command to get the current username.

  • If the user is “anvay”, it proceeds with the subsequent commands.

  • Inside the if block, it attempts to SSH into the tacwin account on the host rpi4.local.

  • The 2> /dev/null redirects the standard error output to /dev/null to suppress any error messages.

  • The output of the SSH command is piped to grep -i “anvay” to search for occurrences of the username “anvay” in the SSH output.

  • The > /dev/null at the end of the pipeline redirects the output of grep to /dev/null to discard it.

  • The if [ “$?” -ne “0” ] condition checks if the exit status of the previous command (grep) is not equal to 0, indicating a failure.

  • If the SSH command fails to find “anvay” in the output, it logs a “login failed” message along with the timestamp to /var/attempt_log.log file.

  • If the SSH command is successful, it logs a “login successful” message along with the timestamp to the same log file.

  • The exit 1 statement terminates the script with an exit code of 1, indicating a failure.

Usage:

  1. Clone the repository:

    git clone https://github.com/asbuch99/FacialRecogProj.git

  2. Install dependencies:

    pip install -r requirements.txt

  3. Train the model:

    python src/facial_recognition.py --train

  4. Run the real-time face recognition application:

    python src/app.py