Facial Recognition for Secured Authentication in the Quantum Era

A facial recognition authentication project supervised by Dr. Khim Chamroeun; First Prize at the 7th Engineering Day Awards 2024.

Overview

A robust facial recognition system for secure authentication in the Quantum Era.

Key Features

  • Facial recognition authentication
  • Machine learning and Deep learning algorithms
  • Secure user verification

Technical Implementation

  • Django backend
  • TensorFlow for deep learning
  • PostgreSQL database
  • Advanced ML algorithms

Model Comparison

LBPH (classical ML) versus InceptionResNetV2 (deep learning), as recorded for the project
MetricLBPH (ML)InceptionResNetV2 (DL)
Real-time accuracy66%98%
Processing speedFast (real-time on CPU)Slower (real-time with GPU acceleration)
Model size33.5 MB298 MB
Training time1 min5 h (Google Colab), 5 min (GPU NVIDIA RTX 4060)
Computational resourcesLow (CPU, 8 GB RAM)High (GPU, 16+ GB RAM)
Hardware requirementsNo GPU required; runs on CPURequires GPU for efficient processing
SecurityModerateHigh

Achievements

  • Awarded 1st Prize at the 7th Engineering Day Awards 2024