Forehead crease patterns improve biometric recognition under occlusion
Balancing Foundation Models and Trait-Specific Learning for Forehead-Crease Recognition
Computer Vision and Pattern Recognition
Summary
Facial recognition often struggles when parts of the face are hidden. This paper shows that the forehead, which is usually visible, has unique crease patterns that can help identify people. The authors compare methods using large language models and specialized face recognition models for recognizing these forehead creases. They find that training models specifically on forehead patterns works better than zero-shot methods that don’t need prior examples. This suggests forehead creases are a useful backup biometric trait when other facial features are obscured.
What this means in practice
- •For security system developers: Use forehead crease recognition to improve identity verification when full facial images are not available.
- •For mobile device manufacturers: Implement robust biometric unlocking that works even when parts of the face are covered by masks or glasses.$Commercial implications: Allows selling enhanced face recognition products that rely on forehead creases as a reliable backup biometric.
Authors
Geetanjali Sharma, Aryan Jain, Dev Nagaich, Aditya Nigam, Raghavendra Ramachandra
Abstract
Biometric recognition often degrades under partial facial vis- ibility, where conventional facial cues are occluded. The forehead region remains consistently visible and contains distinctive crease patterns that can serve as a complementary biometric trait. We present a unified eval- uation of forehead-crease recognition across explainable zero-shot rea- soning, domain-adaptive learning, and trait-specific modeling. FIRE en- ables interpretable zero-shot verification using multimodal large language models, while pretrained face-recognition models are adapted to the fore- head domain through margin-based learning. Under open-set evaluation, the best zero-shot model GPT-4o+FIRE achieves 18.22% EER, while Edge-Forehead achieves 8.30% EER and 54.01% TMR@0.1% FMR. The results show that although MLLMs offer interpretable zero-shot verifi- cation, trait-specific learning remains more effective for robust forehead- crease biometric recognition.