Enhanced Face Authentication With Separate Loss Functions
Anh-Kiet Duong, Hoang-Lan Nguyen, Toan-Thinh Truong
TL;DR
The work tackles mobile face authentication by integrating four modules—face detection, recognition, anti-spoofing, and eye-state classification—into an on-device Android system. It introduces two losses, Large Margin Cotangent Loss ($L_{LMCot}$) and Double Loss, to improve recognition margins and spoof-attack separation, and demonstrates these in both registration and verification workflows. The system uses a 1024-dimensional embedding for recognition, trains with CelebA and CelebA+mask data, and evaluates against related baselines, showing favorable improvements in EER/AUC metrics while maintaining feasible on-device inference times with EfficientNetV2S. Overall, the paper provides a practical, end-to-end framework for secure, attention-aware mobile face authentication with novel margin-based losses and a modular Android implementation.
Abstract
The overall objective of the main project is to propose and develop a system of facial authentication in unlocking phones or applications in phones using facial recognition. The system will include four separate architectures: face detection, face recognition, face spoofing, and classification of closed eyes. In which, we consider the problem of face recognition to be the most important, determining the true identity of the person standing in front of the screen with absolute accuracy is what facial recognition systems need to achieve. Along with the development of the face recognition problem, the problem of the anti-fake face is also gradually becoming popular and equally important. Our goal is to propose and develop two loss functions: LMCot and Double Loss. Then apply them to the face authentication process.