Automatic Face Recognition System Based on Data Augmentation and Transfer Learning

In an era marked by the ubiquity of digital identity verification, the realm of biometrics, especially facial recognition, has been a focal point of research. While it has exhibited impressive performance under controlled conditions, its efficacy degrades under uncontrolled environments. To address this challenge, we present a robust face recognition system inspired by Xception model. We harnessed the power of Data Augmentation and Transfer Learning techniques to proficiently train our proposed Deep CNN model on limited face datasets. These dual techniques not only led to heightened recognition accuracy but also served as robust safeguards against overfitting. Through rigorous testing using the Georgia Tech and ORL benchmark datasets, employing a two-fold cross-validation protocol, our system consistently surpasses contemporary state-of-the-art face recognition methods, achieving an outstanding 100% accuracy rate on both datasats.