Face Detection Based on Deep Learning Approaches: A Comparative Study

Facial recognition is a widely used biometric method for establishing an individual’s identity based on unique and measurable facial characteristics. Human biometrics refers to the statistical analysis of biological measurements unique to individuals. The techniques currently applied in biometrics aim to establish a person’s identity by measuring one of their physical characteristics. Deep learning approaches have significantly advanced the field of facial recognition. This study aims to conduct a comparative analysis of facial recognition approaches based on deep learning. A custom image database adapted to the Moroccan context was collected manually for this purpose. The study encompasses face detection, classification, and facial recognition tasks. Experimental results demonstrate that the MTCNN-ResNet based approach achieves the highest accuracy in both detection and classification. The findings contribute to the understanding of deep learning techniques for facial recognition in the specific context of Morocco.