An Analysis of One-Shot Augmented Learning: A Face Recognition Case Study

Nowadays artificial intelligence models increase in complexity and people tend to doubt how these systems work and how vulnerable they are when they use these systems. The systems indeed use user data, nonetheless, it is possible to reduce this data with techniques like one-shot learning. This study aims to compare how one-shot learning works in face recognition compared to the one-shot augmented learning that uses data augmentation techniques. The research shows that in many face recognition models the data augmentation technique is highly effective. Also, the study allows us to determine the best face recognition model for low data training.