ML Security PapersOn the Impact of Face Segmentation-Based Background Removal on Recognition and Morphing Attack Detection
Eduarda Caldeira 1,2, Guray Ozgur 1,2, Fadi Boutros 1, Naser Damer 1,2
Published on arXiv
2604.20585
Input Manipulation Attack
OWASP ML Top 10 — ML01
Key Finding
Segmentation preprocessing systematically influences both face recognition performance and morphing attack detection, revealing critical tradeoffs for operational biometric systems like EES
This study investigates the impact of face image background correction through segmentation on face recognition and morphing attack detection performance in realistic, unconstrained image capture scenarios. The motivation is driven by operational biometric systems such as the European Entry/Exit System (EES), which require facial enrolment at airports and other border crossing points where controlled backgrounds usually required for such captures cannot always be guaranteed, as well as by accessibility needs that may necessitate image capture outside traditional office environments. By analyzing how such preprocessing steps influence both recognition accuracy and security mechanisms, this work addresses a critical gap between usability-driven image normalization and the reliability requirements of large-scale biometric identification systems. Our study evaluates a comprehensive range of segmentation techniques, three families of morphing attack detection methods, and four distinct face recognition models, using databases that include both controlled and in-the-wild image captures. The results reveal consistent patterns linking segmentation to both recognition performance and face image quality. Additionally, segmentation is shown to systematically influence morphing attack detection performance. These findings highlight the need for careful consideration when deploying such preprocessing techniques in operational biometric systems.
Key Contributions
- Comprehensive evaluation of face segmentation impact on both recognition and morphing attack detection
- Analysis across multiple segmentation techniques, three MAD method families, and four face recognition models
- Reveals systematic influence of segmentation preprocessing on morphing attack detection performance in operational biometric systems
🛡️ Threat Analysis
Morphing attacks are adversarial manipulations of face images designed to evade face recognition systems at inference time—the paper evaluates detection methods against these attacks and analyzes how preprocessing (segmentation) affects attack detection performance.