Improved Facial Recognition for Low-Quality Images Using Image Processing Techniques and Deep Learning Approaches

Facial recognition is a biometric technology used to identify or verify individuals based on their facial features. It involves detecting key facial landmarks, extracting unique features, and comparing them with a stored database. Modern approaches often use deep learning algorithms, particularly Convolutional Neural Networks (CNNs) to achieve high accuracy and robustness in various conditions. This study explores the application of Image processing-based algorithms such as Convolutional Neural Networks (CNNs), used for tasks like face detection, feature extraction, and recognition by automatically learning spatial hierarchies of facial features and Generative Adversarial Networks (GANs), also used for image enhancement tasks like super-resolution and denoising, improving the quality of facial images, especially in low-quality or blurred conditions. Therefore, this research study used the low-quality images and applied on the CNN algorithm against GAN algorithm demonstrate and achieved an accuracy 94.82% and, significantly outperforming against CNN with 88.37% of accuracy. The findings highlight the potential of deep learning approaches in supporting low quality images with improved facial recognition used for biometric verification at border crossings, speeding up the identification process and improving security.