Ning Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ning Li has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Ajian Liu
Institute of Automation, Chinese Academy of Sciences, Beijing, China
5 shared publications - Yanyan Liang
School of Computer Science and Engineering, Macau University of Science and Technology, Macau, China
5 shared publications - Hui Ma
Great Bay University, Dongguan, Guangdong, China
3 shared publications - Yongze Li
School of Computer Science and Engineering, Faculty of Innovation Engineering, Macau University of Science and Technology, Macau, China
3 shared publications - Zhen Lei
Chinese Academy of Sciences
3 shared publications - Chaohao Jiang
School of Computer Science and Engineering, Faculty of Innovation Engineering, Macau University of Science and Technology, Macau, China
2 shared publications
Latest Publications
- Top Cited Papers
- 7 March 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 340
Context-aware knowledge distillation for anomaly detection
- 7 March 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 340
Top Cited PapersContext-aware knowledge distillation for anomaly detection
- 28 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 8167-8177
FA3-CLIP: Frequency-Aware Cues Fusion and Attack-Agnostic Prompt Learning for Unified Face Attack Detection
Facial recognition systems are vulnerable to physical (e.g., printed photos) and digital (e.g., DeepFake) face attacks. Existing methods struggle to simultaneously detect physical and digital attacks due to: 1) significant intra-class variations between these attack types, and 2) the inadequacy of spatial information alone to comprehensively capture live and fake cues. To address these issues, we propose a unified attack detection model termed Frequency-Aware and Attack-Agnostic CLIP (FA3-CLIP), which introduces attack-agnostic prompt learning to express generic live and fake cues derived from the fusion of spatial and frequency features, enabling unified detection of live faces and all categories of attacks. Specifically, the attack-agnostic prompt module generates generic live and fake prompts within the language branch to extract corresponding generic representations from both live and fake faces, guiding the model to learn a unified feature space for unified attack detection. Meanwhile, the module adaptively generates the live/fake conditional bias from the original spatial and frequency information to optimize the generic prompts accordingly, reducing the impact of intra-class variations. We further propose a dual-stream cues fusion framework in the vision branch, which leverages frequency information to complement subtle cues that are difficult to capture in the spatial domain. In addition, a frequency compression block is utilized in the frequency stream, which reduces redundancy in frequency features while preserving the diversity of crucial cues. We also establish new challenging protocols to facilitate unified face attack detection effectiveness. Experimental results on multiple benchmarks demonstrate that FA3-CLIP significantly improves performance, reducing ACER by over 1.2% on UniAttackData, and increasing AUC by more than 3% as well as reducing EER by over 4% on the JFSFDB dataset. The source code is available at https://github.com/YongzeLi/FA3-CLIP
- 28 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 8167-8177
FA3-CLIP: Frequency-Aware Cues Fusion and Attack-Agnostic Prompt Learning for Unified Face Attack Detection
Facial recognition systems are vulnerable to physical (e.g., printed photos) and digital (e.g., DeepFake) face attacks. Existing methods struggle to simultaneously detect physical and digital attacks due to: 1) significant intra-class variations between these attack types, and 2) the inadequacy of spatial information alone to comprehensively capture live and fake cues. To address these issues, we propose a unified attack detection model termed Frequency-Aware and Attack-Agnostic CLIP (FA3-CLIP), which introduces attack-agnostic prompt learning to express generic live and fake cues derived from the fusion of spatial and frequency features, enabling unified detection of live faces and all categories of attacks. Specifically, the attack-agnostic prompt module generates generic live and fake prompts within the language branch to extract corresponding generic representations from both live and fake faces, guiding the model to learn a unified feature space for unified attack detection. Meanwhile, the module adaptively generates the live/fake conditional bias from the original spatial and frequency information to optimize the generic prompts accordingly, reducing the impact of intra-class variations. We further propose a dual-stream cues fusion framework in the vision branch, which leverages frequency information to complement subtle cues that are difficult to capture in the spatial domain. In addition, a frequency compression block is utilized in the frequency stream, which reduces redundancy in frequency features while preserving the diversity of crucial cues. We also establish new challenging protocols to facilitate unified face attack detection effectiveness. Experimental results on multiple benchmarks demonstrate that FA3-CLIP significantly improves performance, reducing ACER by over 1.2% on UniAttackData, and increasing AUC by more than 3% as well as reducing EER by over 4% on the JFSFDB dataset. The source code is available at https://github.com/YongzeLi/FA3-CLIP
- 11 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. PP (15513203) , 1-12
Knowledge Distillation-Based Anomaly Detection via Adaptive Discrepancy Optimization
Knowledge distillation has emerged as a primary solution for anomaly detection, leveraging feature discrepancies between teacher–student (T–S) networks to locate anomalies. However, previous approaches suffer from ambiguous feature discrepancies, which hinder effective anomaly detection due to two main challenges: 1) overgeneralization, where the student network excessively mimics teacher features in anomalous regions, and 2) semantic bias between T–S networks in normal regions. To address these issues, we propose an Adaptive Discrepancy Optimization (Ado) block. The Ado block adaptively calibrates feature discrepancies by reducing overgeneralization in anomalous regions and selectively aligning semantic features in normal regions via learnable feature offsets. This versatile block can be seamlessly integrated into various distillation-based methods. Experimental results demonstrate that the Ado block significantly enhances performance across 11 different knowledge distillation frameworks on two widely used datasets. Notably, when integrated with the Ado block, RD4AD achieves a 22% relative improvement in pixel-level PRO on the VisA dataset. In addition, a real-world keyboard inspection application further validates the effectiveness of the Ado block.
- 11 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. PP (15513203) , 1-12
Knowledge Distillation-Based Anomaly Detection via Adaptive Discrepancy Optimization
Knowledge distillation has emerged as a primary solution for anomaly detection, leveraging feature discrepancies between teacher–student (T–S) networks to locate anomalies. However, previous approaches suffer from ambiguous feature discrepancies, which hinder effective anomaly detection due to two main challenges: 1) overgeneralization, where the student network excessively mimics teacher features in anomalous regions, and 2) semantic bias between T–S networks in normal regions. To address these issues, we propose an Adaptive Discrepancy Optimization (Ado) block. The Ado block adaptively calibrates feature discrepancies by reducing overgeneralization in anomalous regions and selectively aligning semantic features in normal regions via learnable feature offsets. This versatile block can be seamlessly integrated into various distillation-based methods. Experimental results demonstrate that the Ado block significantly enhances performance across 11 different knowledge distillation frameworks on two widely used datasets. Notably, when integrated with the Ado block, RD4AD achieves a 22% relative improvement in pixel-level PRO on the VisA dataset. In addition, a real-world keyboard inspection application further validates the effectiveness of the Ado block.
- Open Access
- 1 January 2025
- Published by Institution of Engineering and Technology (IET) in IET Biometrics
- Vol. 2025 (1)
Layer‐Wise Cue Alignment for Unified Face Attack Detection With Vision‐Language Model
Unified face attack detection (UAD), which simultaneously addresses physical presentation attacks (PAs) and digital forgery attacks (DAs) with a vision‐language model, remains challenging due to the difficulty of effectively separating live and fake cues. The challenges mainly arise from two aspects: (1) text prompts are insufficiently aligned with visual features across layers, and (2) patch tokens containing live and fake cues often overlap, leading to ambiguous attribution and small decision margins. To address these problems, we propose a novel Layer‐wise Cue Alignment framework (LCA) that leverages textual features to extract both layer‐wise and global cues from patch tokens, and we further introduce a new training strategy to improve the separation of live and fake cues. Specifically, the layer‐wise prompts are obtained by the cue matching block (CMB), which matches textual features with patch embeddings at each transformer layer, and the layer‐level cues are injected into the visual features of each layer and further aggregated by the cue fusion block (CFB) to form comprehensive prompts that enhance the overall visual representation. Moreover, we design a complementary supervision mechanism (CSM) that suppresses forgery cues in live faces while enforcing mutual exclusivity between live and fake cues in attack samples to improve the reliability of cue separation. Extensive experiments on multiple benchmarks demonstrate that our framework achieves state‐of‐the‐art performance on most protocols of the datasets.
- 1 January 2025
- Published by Institution of Engineering and Technology (IET) in IET Biometrics
- Vol. 2025 (1)
Open AccessLayer‐Wise Cue Alignment for Unified Face Attack Detection With Vision‐Language Model
Unified face attack detection (UAD), which simultaneously addresses physical presentation attacks (PAs) and digital forgery attacks (DAs) with a vision‐language model, remains challenging due to the difficulty of effectively separating live and fake cues. The challenges mainly arise from two aspects: (1) text prompts are insufficiently aligned with visual features across layers, and (2) patch tokens containing live and fake cues often overlap, leading to ambiguous attribution and small decision margins. To address these problems, we propose a novel Layer‐wise Cue Alignment framework (LCA) that leverages textual features to extract both layer‐wise and global cues from patch tokens, and we further introduce a new training strategy to improve the separation of live and fake cues. Specifically, the layer‐wise prompts are obtained by the cue matching block (CMB), which matches textual features with patch embeddings at each transformer layer, and the layer‐level cues are injected into the visual features of each layer and further aggregated by the cue fusion block (CFB) to form comprehensive prompts that enhance the overall visual representation. Moreover, we design a complementary supervision mechanism (CSM) that suppresses forgery cues in live faces while enforcing mutual exclusivity between live and fake cues in attack samples to improve the reliability of cue separation. Extensive experiments on multiple benchmarks demonstrate that our framework achieves state‐of‐the‐art performance on most protocols of the datasets.
- Open Access
- 1 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 174910-174921
Dual-Teacher Guided Denoising Distillation for Anomaly Detection
- 1 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 174910-174921
Open AccessDual-Teacher Guided Denoising Distillation for Anomaly Detection