On the Robustness of Face Recognition Algorithms Against Attacks and Bias

Face recognition algorithms have demonstrated very high recognition\nperformance, suggesting suitability for real world applications. Despite the\nenhanced accuracies, robustness of these algorithms against attacks and bias\nhas been challenged. This paper summarizes different ways in which the\nrobustness of a face recognition algorithm is challenged, which can severely\naffect its intended working. Different types of attacks such as physical\npresentation attacks, disguise/makeup, digital adversarial attacks, and\nmorphing/tampering using GANs have been discussed. We also present a discussion\non the effect of bias on face recognition models and showcase that factors such\nas age and gender variations affect the performance of modern algorithms. The\npaper also presents the potential reasons for these challenges and some of the\nfuture research directions for increasing the robustness of face recognition\nmodels.\n