2026· Computers, Materials & Continua· Vol 89, pp. 1-10· 0 citations· 52 references
TL;DR
The Semantic-Anchored Test-Time Domain Generalization (SA-TTDG) framework is introduced, introducing a Text-Anchored Style Projection (TASP), which utilizes rich linguistic priors from Vision-Language Models (VLMs) to initialize and strongly constrain learnable style bases.
Abstract
: To ensure the reliability of biometric authentication, Face Anti-Spoofing (FAS) models must accurately detect presentation attacks. However, due to the highly complex distribution shifts caused by variations in style, cross-domain generalization remains a significant challenge. Test-Time Domain Generalization (TTDG) has recently surfaced as an innovative framework, facilitating the adaptation of unseen samples to source-domain characteristics through the strategic utilization of learned style bases. Nevertheless, existing TTDG methods optimize randomly initialized style bases solely through statistical objectives, leaving a critical research gap: the lack of explicit semantic constraints inevitably leads to hierarchical semantic inconsistency and weakens subtle spoofing cues. To fill this gap, we introduce the Semantic-Anchored Test-Time Domain Generalization (SA-TTDG) framework. The core novelty of our approach lies in introducing a Text-Anchored Style Projection (TASP), which utilizes rich linguistic priors from Vision-Language Models (VLMs) to initialize and strongly constrain learnable style bases. By anchoring these bases to explicit semantic concepts, TASP encourages semantic consistency across hierarchical feature representations. Furthermore, to fully exploit these semantically aligned style representations, we design a Semantic Prompt Modulation (SPM) module driven by a Style-Query Cross-Attention (SQ-CA) mechanism. Instead of using static queries, SPM dynamically retrieves multi-level style cues to generate domain-sensitive prompts, which effectively modulate the original visual features. This process helps enhance subtle spoofing-related cues while preserving the underlying content structure. Evaluations under standard leave-one-domain-out protocols demonstrate that the proposed framework consistently reduces cross-domain classification errors compared to existing statistical TTDG baselines.
Facial biometric recognition systems currently face compound threats intertwining generative AI and high-fidelity physical spoofing. Existing defenses suffer from systemic bottlenecks, including poor generalization, non-auditable reasoning, and reliance on massive, low-quality datasets. To address these challenges, we propose Multimodal Large Language Models (MFAD) for face anti-spoofing detection, an explainable reasoning system for Unified Face Anti-Spoofing Detection (UFAD), accompanied by a semantic-level annotation benchmark. Unlike methods relying on external tools or coarse alignment, MFAD activates the intrinsic reasoning capabilities of Multimodal Large Language Models (MLLMs) via a fine-grained pixel-semantic anchoring mechanism. This eliminates localization hallucinations and ensures auditable reasoning paths. We introduce a cross-attack semantic-level unified annotation paradigm: by annotating only 1,000 precise masks per attack category, we generate reasoning evidence chains strictly corresponding to spoofed regions. Supervised fine-tuning on the Qwen-VL foundation model demonstrates that, using limited high-quality samples, the system achieves a 40-50% relative reduction in in-domain ACER and restricts cross-domain performance degradation to within 11.62%/5.23%, significantly outperforming existing frameworks. Furthermore, under white-box adversarial attacks, detection accuracy drops by only 3.2%, validating the robustness of semantic anchoring compared to models trained on massive short-text data. Domain practitioners rated the evidence reliability of reasoning paths at 4.57/5, with inference latency satisfying real-time deployment requirements. These results confirm that a few-shot, high-quality semantic annotation paradigm is effective for building trustworthy, explainable, and cost-efficient UFAD systems.
Xiaoyong Yu, Rongzhen Li, Shuming Shi et al.· 0 citations
Face recognition has been widely applied in identity authentication systems, but its vulnerability to various presentation attacks poses significant security risks. As a result, face anti-spoofing (FAS) has become one of the key technologies for ensuring the reliability of such systems. Most existing domain- generalized FAS (DGFAS) approaches rely on domain distribution alignment to learn cross-domain invariant representations. However, many of these methods independently model each sample while overlooking semantic consistency among cross-domain samples and different regions within each sample, making them vulnerable to semantic drift. Moreover, some existing methods introduce additional pixel-level supervision signals, such as pseudo-depth maps and binary masks, which increase annotation costs and limit their generalizability. To address these challenges, we propose DSCM-FAS, a novel framework that combines inter-sample and intra-sample semantic consistency to enhance robustness and cross-domain generalization, without requiring auxiliary supervision. Specifically, we design a Dual Semantic Consistency Module (DSCM). Across samples, contrastive learning is leveraged to learn discriminative representations, followed by the adaptive construction of a high-confidence similarity adjacency graph via statistical thresholding. The resulting graph is then fed into a graph convolutional network (GCN) to strengthen cross-domain semantic consistency, thereby improving the model's robustness to domain shifts. Within each sample, the intermediate embedding features are first partitioned into multiple patches, and Laplacian Regularization is introduced to constrain the semantic relationships among different patches. This effectively suppresses local noise interference and promotes the learning of more robust and semantically consistent feature representations. Extensive experiments on four public FAS datasets demonstrate that DSCM-FAS consistently outperforms state-of-the-art methods under various protocols. These results validate the effectiveness of our DSCM-FAS in improving the cross-domain generalization of FAS models.
Hui Xing, Siok Yee Tan, Faizan Qamar et al.· Journal of King Saud Univers...· 0 citations
Face anti-spoofing (FAS) is increasingly expected to provide not only bona fide/spoof decisions, but also attack semantics and image-grounded evidence for human inspection. Existing discriminative FAS models remain largely label-centric, while recent MLLM-based methods offer structured outputs but still rely mainly on supervised fine-tuning, often producing template-like rationales and weak optimization for difficult attacks. We propose FAS-R1, a two-stage reasoning-oriented MLLM framework for unified FAS prediction, covering authenticity classification, attack-type recognition and spoof-region localization. FAS-R1 first uses FAS-R1-23K, a high-quality long-CoT dataset, for cold-start supervised fine-tuning, and then performs FAS-specific GRPO post-training. Degradation-Simulated Augmentation (DSA) encourages stable spoof-cue reasoning across visual-quality shifts, while Difficulty-Aware GRPO (DA-GRPO) mitigates easy-sample dominance that may leave difficult task--attack groups under-optimized, especially for subtle or ambiguous attacks such as makeup and mask attacks. The main 3B FAS-R1 model achieves 98.75\% authenticity accuracy, 93.33\% attack-type accuracy, and 96.30/94.73\% AP@40/AP@50 in-domain. It also outperforms the compared systems in cross-domain authenticity generalization and answer-and-rationale quality. Experiments with different base models further show favorable scaling behavior. The code will be released soon.
Hongyang Wang, Yichen Shi, Hongrui Li et al.· 0 citations
Open-world face anti-spoofing must address both covariate and semantic shifts: source and target domains differ in imaging conditions, while target domains contain diverse attack types absent from training. Existing prompt-based approaches often express spoofing through category semantics or language guidance, which is effective for modeling high-level concepts but is less suited to explicitly capturing the evolving fine-grained and spatially heterogeneous forensic evidence of unseen attacks. Motivated by the hypothesis that many unseen attacks can be characterized by new combinations of recurring visual cues, we propose a compositional forensic visual prompt learning framework that operates entirely in the visual feature space. Built on a frozen ViT-based vision foundation model, the framework employs patch-aware attention to refine a shared set of learnable micro-forensic primitives into localized forensic evidence units derived from image patches. Class-specific global contextual prompts then provide input-dependent routing weights that adaptively select and compose these primitives into compositional forensic visual prompts for real/spoof discrimination. The primitives are not assigned predefined semantic meanings; instead, their specialization and reuse emerge from shared parameterization and joint optimization across categories. Extensive experiments on nine open-world protocols demonstrate state-of-the-art performance, strong cross-domain generalization, and robust adaptation to unseen attacks.
Fangling Jiang, Qi Li, Bing Liu et al.· 0 citations
Face recognition systems are increasingly deployed in security-critical applications, yet they remain vulnerable to presentation and morph attacks. Presentation attack detection (PAD) and morphing attack detection (MAD) are therefore essential components of trustworthy face biometrics. Despite advancements in PAD and MAD methods, existing detectors suffer from limited generalization and degrade in cross-dataset evaluation. In this paper, we systematically investigate whether general-purpose foundation models (FMs) and multimodal large language models (MLLMs) encode PAD-relevant and MAD-relevant information, and how such models can best be deployed for both tasks. We study five approaches with increasing access to the internal information of the model: (i) zero-shot prompting of off-the-shelf MLLMs; (ii) training a shallow model on the next-token logit probabilities at the output of the MLLM; (iii) parameter-efficient fine-tuning on task-specific question-answer data, yielding two specialized MLLMs, called PADLLM and MADLLM, which additionally provide textual reasoning for their decisions; (iv) linear probing of frozen vision encoders; and (v) fine-tuning of vision encoders of FMs and MLLMs. We benchmark 16 open-weight MLLMs and 30 vision encoder backbones on four PAD datasets (MSU-MFSD, CASIA-FASD, Replay-Attack, and OULU-NPU) and four MAD datasets (FFHQ, FRGC, FRLL, and FERET). Our experiments show that FMs and MLLMs can achieve significant performance for PAD and MAD. In addition, the fine-tuned models achieve state-of-the-art detection performance in cross-dataset evaluation, indicating that general-purpose pretrained representations carry substantial attack-relevant information. Source code of all our experiments will be publicly released.
Hatef Otroshi Shahreza, Asif Hussain Khan, P. Lorenz et al.· 0 citations
While foundation models have significantly advanced human recognition across diverse modalities, they predominantly rely on static, geometric feature extraction. This approach fundamentally diverges from human perception. Consequently, current models often suffer from"semantic blindness,"overfitting to transient noise while failing to leverage invariant soft biometrics, and struggle to capture temporal motion signatures. To bridge this gap, we propose SapiensID 2.0, a human recognition framework enriched with both semantic and temporal awareness. To overcome the lack of soft-biometric annotations, we transfer zero-shot semantic knowledge from Multimodal Large Language Models (MLLMs) into a discriminative embedding space. We resolve the dimensional mismatch between these spaces using Invariant Trait Alignment (ITA) to distill core persistent traits, and Transient Noise Disentanglement (TND) to decouple artifacts like clothing. Furthermore, we design a Kinematic Semantic Attention Head (K-SAH) that extends spatial attention across temporal windows. By tracking semantic patches over time, K-SAH captures rich kinematic signatures without requiring large-scale video datasets. Extensive experiments demonstrate that SapiensID 2.0 achieves state-of-the-art performance across image- and video-based person re-identification and gait recognition, while maintaining robust face recognition capabilities.
Yiyang Su, Jie Zhu, Feng Liu et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
