SRL-MAD: Structured Residual Latents for One-Class Morphing Attack Detection

2026-03-16

AI Summary

This work addresses the challenging one-class detection problem in face morphing attack detection (MAD), where models are trained exclusively on bona fide facial images yet must generalize to unseen morphing attacks. To this end, the authors propose a novel approach based on structured residual spectral representations. Departing from conventional reconstruction-error paradigms, the method preserves the two-dimensional spectral structure in the Fourier domain and introduces learnable annular spectral projections combined with inductive biases that explicitly segment and model interactions across low-, mid-, and high-frequency bands. Discriminative scores are generated directly in the latent space without relying on reconstruction. Extensive experiments demonstrate that the proposed method significantly outperforms both existing one-class and supervised MAD approaches on the FERET-Morph, FRLL-Morph, and MorDIFF datasets, confirming the efficacy and superiority of frequency-aware structured representations for open-set morphing attack detection.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Multimodal Learning
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Face morphing attacks represent a significant threat to biometric systems as they allow multiple identities to be combined into a single face. While supervised morphing attack detection (MAD) methods have shown promising performance, their reliance on attack-labeled data limits generalization to unseen morphing attacks. This has motivated increasing interest in one-class MAD, where models are trained exclusively on bona fide samples and are expected to detect unseen attacks as deviations from the normal facial structure. In this context, we introduce SRL-MAD, a one-class single-image MAD that uses structured residual Fourier representations for open-set morphing attack detection. Starting from a residual frequency map that suppresses image-specific spectral trends, we preserve the two-dimensional organization of the Fourier domain through a ring-based representation and replace azimuthal averaging with a learnable ring-wise spectral projection. To further encode domain knowledge about where morphing artifacts arise, we impose a frequency-informed inductive bias by organizing spectral evidence into low, mid, and high-frequency bands and learning cross-band interactions. These structured spectral features are mapped into a latent space designed for direct scoring, avoiding the reliance on reconstruction errors. Extensive evaluation on FERET-Morph, FRLL-Morph, and MorDIFF demonstrates that SRL-MAD consistently outperforms recent one-class and supervised MAD models. Overall, our results show that learning frequency-aware projections provides a more discriminative alternative to azimuthal spectral summarization for one-class morphing attack detection.

Problem

• morphing attack detection
• one-class classification
• face recognition security
• open-set detection
• biometric spoofing

Innovation

• structured residual latents
• one-class morphing attack detection
• Fourier ring representation
• frequency-aware projection
• open-set MAD