Facial Recognition System Based on Genetic Algorithm Improved ROI-KNN Convolutional Neural Network.

Area of Science:

• Computer Science
• Artificial Intelligence
• Biometrics

Background:

• Facial recognition systems, utilizing AI and biometrics, are widely applied but suffer from low accuracy, especially with uncooperative subjects.
• Existing Region of Interest (ROI)-K nearest neighbor (KNN) convolutional neural network models show limitations in accuracy and security under challenging conditions like poor lighting or occlusion.

Purpose of the Study:

• To improve the accuracy, speed, and security of facial recognition systems.
• To address the limitations of current ROI-KNN convolutional neural network models in real-world scenarios.

Main Methods:

• Integration of adaptive quantum genetic algorithm, improved marker line graph genetic algorithm, and feature weight value genetic algorithm.
• Enhancement of the existing ROI-KNN convolutional neural network framework using genetic algorithms.

Main Results:

• Facial recognition accuracy increased by 4.99%.
• Recognition speed improved by 7.46%.
• Recognition security saw a 2.66% increase.

Conclusions:

• Genetic algorithm-based improvements significantly enhance facial recognition system performance.
• The developed system offers a more robust and secure solution for various applications.