Chun Pong Lau
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Chun Pong Lau has a total of 73 co-authors. Here are the top co-authors by number of shared publications:
- Lok Ming Lui
Department of Mathematics, The Chinese University of Hong Kong, Hong Kong, China
6 shared publications - Basem Shihada
King Abdullah University of Science and Technology (KAUST)
5 shared publications - Jiang Liu
Department of Electrical and Computer Engineering, Johns Hopkins University, Baltimore, MD, USA
3 shared publications Rama Chellappa
Department of Electrical and Computer Engineering and the Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, USA
3 shared publications- Rama Chellappa
Johns Hopkins University
3 shared publications - Zhongliang Guo
School of Computer Science, University of St Andrews, St Andrews, U.K.
3 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 7410-7419
StyleCom: Taming Style Transfer and Collage Prior for Training-Based Cross-Domain Image Composition
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 7410-7419
Open AccessStyleCom: Taming Style Transfer and Collage Prior for Training-Based Cross-Domain Image Composition
- Open Access
Global Variation in Timing of Allergenic Food Introduction for Food Allergy Prevention: An International Survey of Healthcare Professionals
Background: IgE‐mediated food allergy prevention guidelines are now established in many countries, but their implementation in routine clinical practice globally remains unclear. Methods: An anonymous online survey was distributed to healthcare professionals (HCPs) through the World Allergy Organization network between December 2024 and April 2025, collecting data on food allergy prevention recommendations in daily clinical practice. Results: The analysis included 731 healthcare professionals from 80 countries worldwide: Asia (36.7%), Europe (27.1%), North America (15.9%), and Rest of the World (20.0%; Russia, South America, Africa, Oceania). Unsupervised clustering revealed two distinct practice patterns—one favoring early allergen introduction and one favoring later introduction—with timing varying by allergen type. For peanuts, clusters diverged between early (median 6 months of age) and late (median 18 months of age) introduction recommendations for high‐risk infants. Allergen introduction timing was independently driven by region and specialty: North American HCPs recommend peanut introduction 6.7 months earlier than those in Asia (5.48 3.32 vs. 12.18 7.27 months; p < 0.001), and pediatric allergists globally advocate for introduction 2.37 months earlier than non‐allergists (8.72 5.85 vs. 11.09 7.14 months, p = 0.024). This disparity is more pronounced in high‐risk infants than in normal‐risk infants. Conclusions: The variability in allergen introduction practices, driven by geography and medical specialty, highlights a persistent gap in prevention guidelines adoption. This divergence likely reflects both suboptimal implementation of existing recommendations and regional differences in food allergy epidemiology. These findings underscore the need for population‐tailored allergen introduction strategies across diverse healthcare settings.
Open AccessGlobal Variation in Timing of Allergenic Food Introduction for Food Allergy Prevention: An International Survey of Healthcare Professionals
Background: IgE‐mediated food allergy prevention guidelines are now established in many countries, but their implementation in routine clinical practice globally remains unclear. Methods: An anonymous online survey was distributed to healthcare professionals (HCPs) through the World Allergy Organization network between December 2024 and April 2025, collecting data on food allergy prevention recommendations in daily clinical practice. Results: The analysis included 731 healthcare professionals from 80 countries worldwide: Asia (36.7%), Europe (27.1%), North America (15.9%), and Rest of the World (20.0%; Russia, South America, Africa, Oceania). Unsupervised clustering revealed two distinct practice patterns—one favoring early allergen introduction and one favoring later introduction—with timing varying by allergen type. For peanuts, clusters diverged between early (median 6 months of age) and late (median 18 months of age) introduction recommendations for high‐risk infants. Allergen introduction timing was independently driven by region and specialty: North American HCPs recommend peanut introduction 6.7 months earlier than those in Asia (5.48 3.32 vs. 12.18 7.27 months; p < 0.001), and pediatric allergists globally advocate for introduction 2.37 months earlier than non‐allergists (8.72 5.85 vs. 11.09 7.14 months, p = 0.024). This disparity is more pronounced in high‐risk infants than in normal‐risk infants. Conclusions: The variability in allergen introduction practices, driven by geography and medical specialty, highlights a persistent gap in prevention guidelines adoption. This divergence likely reflects both suboptimal implementation of existing recommendations and regional differences in food allergy epidemiology. These findings underscore the need for population‐tailored allergen introduction strategies across diverse healthcare settings.
- 19 June 2026
- Published by Elsevier in Pattern Recognition
- Vol. 180
Segment and inpaint: Defending object detectors with a unified segment and inpaint paradigm
- 19 June 2026
- Published by Elsevier in Pattern Recognition
- Vol. 180
Segment and inpaint: Defending object detectors with a unified segment and inpaint paradigm
- 3 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 12918-12933
A Gray-Box Attack Against Latent Diffusion Model-Based Image Editing by Posterior Collapse
Recent advancements in Latent Diffusion Models (LDMs) have revolutionized image synthesis and manipulation, raising significant concerns about data misappropriation and intellectual property infringement. While adversarial attacks have been extensively explored as a protective measure against such misuse of generative AI, current approaches are severely limited by their heavy reliance on model-specific knowledge and substantial computational costs. Drawing inspiration from the posterior collapse phenomenon observed in VAE training, we propose the Posterior Collapse Attack (PCA), a novel framework for protecting images from unauthorized manipulation. Through comprehensive theoretical analysis and empirical validation, we identify two distinct collapse phenomena during VAE inference: diffusion collapse and concentration collapse. Based on this discovery, we design a unified loss function that can flexibly achieve both types of collapse through parameter adjustment, each corresponding to different protection objectives in preventing image manipulation. Our method significantly reduces dependence on model-specific knowledge by requiring access to only the VAE encoder, which constitutes less than 4% of LDM parameters. Notably, PCA achieves prompt-invariant protection by operating on the VAE encoder before text conditioning occurs, eliminating the need for empty prompt optimization required by existing methods. This minimal requirement enables PCA to maintain adequate transferability across various VAE-based LDM architectures while effectively preventing unauthorized image editing. Extensive experiments show PCA outperforms existing techniques in protection effectiveness, computational efficiency (runtime and VRAM), and generalization across VAE-based LDM variants. Our code is available at https://github.com/ZhongliangGuo/PosteriorCollapseAttack.
- 3 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 12918-12933
A Gray-Box Attack Against Latent Diffusion Model-Based Image Editing by Posterior Collapse
Recent advancements in Latent Diffusion Models (LDMs) have revolutionized image synthesis and manipulation, raising significant concerns about data misappropriation and intellectual property infringement. While adversarial attacks have been extensively explored as a protective measure against such misuse of generative AI, current approaches are severely limited by their heavy reliance on model-specific knowledge and substantial computational costs. Drawing inspiration from the posterior collapse phenomenon observed in VAE training, we propose the Posterior Collapse Attack (PCA), a novel framework for protecting images from unauthorized manipulation. Through comprehensive theoretical analysis and empirical validation, we identify two distinct collapse phenomena during VAE inference: diffusion collapse and concentration collapse. Based on this discovery, we design a unified loss function that can flexibly achieve both types of collapse through parameter adjustment, each corresponding to different protection objectives in preventing image manipulation. Our method significantly reduces dependence on model-specific knowledge by requiring access to only the VAE encoder, which constitutes less than 4% of LDM parameters. Notably, PCA achieves prompt-invariant protection by operating on the VAE encoder before text conditioning occurs, eliminating the need for empty prompt optimization required by existing methods. This minimal requirement enables PCA to maintain adequate transferability across various VAE-based LDM architectures while effectively preventing unauthorized image editing. Extensive experiments show PCA outperforms existing techniques in protection effectiveness, computational efficiency (runtime and VRAM), and generalization across VAE-based LDM variants. Our code is available at https://github.com/ZhongliangGuo/PosteriorCollapseAttack.
- Top Cited PapersOpen Access
- 24 November 2025
- Published by Elsevier in Pattern Recognition
- Vol. 173
DiffProtect: Generative adversarial examples using diffusion models for facial privacy protection
- 24 November 2025
- Published by Elsevier in Pattern Recognition
- Vol. 173
Top Cited PapersOpen AccessDiffProtect: Generative adversarial examples using diffusion models for facial privacy protection
- Open Access
- 4 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 7 (4) , 940-952
Distillation-Guided Representation Learning for Unconstrained Video Human Authentication
Human authentication is an important and challenging biometric task, particularly from unconstrained videos. While body recognition is a popular approach, gait recognition holds the promise of robustly identifying subjects based on walking patterns instead of appearance information. Previous gait-based approaches have performed well for curated indoor scenes; however, they tend to underperform in unconstrained situations. To address these challenges, we propose a framework, termed Holistic GAit DEtection and Recognition (H-GADER), for human authentication in challenging outdoor scenarios. Specifically, H-GADER leverages a Double Helical Signature to detect segments that contain human movement and builds discriminative features through a novel gait recognition method. To further enhance robustness, H-GADER encodes viewpoint information in its architecture, and distills learned representations from an auxiliary RGB recognition model; this allows H-GADER to learn from maximum amount of data at training time. At test time, H-GADER infers solely from the silhouette modality. Furthermore, we introduce a body recognition model through semantic, large-scale, self-supervised training to complement gait recognition. By conditionally fusing gait and body representations based on the presence/absence of gait information as decided by the gait detection, we demonstrate significant improvements compared to when a single modality or a naive feature ensemble is used. We evaluate our method on multiple existing State-of-The-Arts(SoTA) gait baselines and demonstrate consistent improvements on indoor and outdoor datasets, especially on the BRIAR dataset, which features unconstrained, long-distance videos, achieving a 28.9% improvement.
- 4 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 7 (4) , 940-952
Open AccessDistillation-Guided Representation Learning for Unconstrained Video Human Authentication
Human authentication is an important and challenging biometric task, particularly from unconstrained videos. While body recognition is a popular approach, gait recognition holds the promise of robustly identifying subjects based on walking patterns instead of appearance information. Previous gait-based approaches have performed well for curated indoor scenes; however, they tend to underperform in unconstrained situations. To address these challenges, we propose a framework, termed Holistic GAit DEtection and Recognition (H-GADER), for human authentication in challenging outdoor scenarios. Specifically, H-GADER leverages a Double Helical Signature to detect segments that contain human movement and builds discriminative features through a novel gait recognition method. To further enhance robustness, H-GADER encodes viewpoint information in its architecture, and distills learned representations from an auxiliary RGB recognition model; this allows H-GADER to learn from maximum amount of data at training time. At test time, H-GADER infers solely from the silhouette modality. Furthermore, we introduce a body recognition model through semantic, large-scale, self-supervised training to complement gait recognition. By conditionally fusing gait and body representations based on the presence/absence of gait information as decided by the gait detection, we demonstrate significant improvements compared to when a single modality or a naive feature ensemble is used. We evaluate our method on multiple existing State-of-The-Arts(SoTA) gait baselines and demonstrate consistent improvements on indoor and outdoor datasets, especially on the BRIAR dataset, which features unconstrained, long-distance videos, achieving a 28.9% improvement.
- 8 July 2025
- Published by Elsevier in Pattern Recognition
- Vol. 171
Artwork protection against unauthorized neural style transfer and aesthetic color distance metric
- 8 July 2025
- Published by Elsevier in Pattern Recognition
- Vol. 171
Artwork protection against unauthorized neural style transfer and aesthetic color distance metric
- 19 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. PP (01628828) , 1-14
Interpolated Joint Space Adversarial Training for Robust and Generalizable Defenses
Adversarial training (AT) is considered to be one of the most reliable defenses against adversarial attacks. However, models trained with AT sacrifice standard accuracy and do not generalize well to unseen attacks. Some examples of recent works show generalization improvement with adversarial samples under unseen threat models are, on-manifold threat model or neural perceptual threat model. However, the former requires exact manifold information while the latter requires algorithm relaxation. Motivated by these considerations, we propose a novel threat model called Joint Space Threat Model (JSTM), which exploits the underlying manifold information with Normalizing Flow, ensuring that the exact manifold assumption holds. Under JSTM, we develop novel adversarial attacks and defenses. Specifically, we propose the Robust Mixup strategy in which we maximize the adversity of the interpolated images and gain robustness and prevent overfitting. Our experiments show that Interpolated Joint Space Adversarial Training (IJSAT) achieves good performance in standard accuracy, robustness, and generalization. IJSAT is also flexible and can be used as a data augmentation method to improve standard accuracy and combined with many existing AT approaches can improve robustness. We demonstrate the effectiveness of our approach on three benchmark datasets, CIFAR-10/100, OM-ImageNet and CIFAR-10-C.
- 19 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. PP (01628828) , 1-14
Interpolated Joint Space Adversarial Training for Robust and Generalizable Defenses
Adversarial training (AT) is considered to be one of the most reliable defenses against adversarial attacks. However, models trained with AT sacrifice standard accuracy and do not generalize well to unseen attacks. Some examples of recent works show generalization improvement with adversarial samples under unseen threat models are, on-manifold threat model or neural perceptual threat model. However, the former requires exact manifold information while the latter requires algorithm relaxation. Motivated by these considerations, we propose a novel threat model called Joint Space Threat Model (JSTM), which exploits the underlying manifold information with Normalizing Flow, ensuring that the exact manifold assumption holds. Under JSTM, we develop novel adversarial attacks and defenses. Specifically, we propose the Robust Mixup strategy in which we maximize the adversity of the interpolated images and gain robustness and prevent overfitting. Our experiments show that Interpolated Joint Space Adversarial Training (IJSAT) achieves good performance in standard accuracy, robustness, and generalization. IJSAT is also flexible and can be used as a data augmentation method to improve standard accuracy and combined with many existing AT approaches can improve robustness. We demonstrate the effectiveness of our approach on three benchmark datasets, CIFAR-10/100, OM-ImageNet and CIFAR-10-C.
- 17 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 17 (15566013) , 2364-2377
Mutual Adversarial Training: Learning Together is Better Than Going Alone
Recent studies have shown that robustness to adversarial attacks can be transferred across deep neural networks. In other words, we can make a weak model more robust with the help of a strong teacher model. In this paper, we ask if models can “learn together” and “teach each other” to achieve better robustness instead of learning from a static teacher. We study how interactions among models enhance robustness via knowledge distillation. We propose mutual adversarial training (MAT), in which multiple models are trained together and share the knowledge of adversarial examples to achieve improved robustness. MAT allows robust models to explore a larger space of adversarial samples and find more robust feature spaces and decision boundaries. Through extensive experiments on the CIFAR-10, CIFAR-100, and mini-ImageNet datasets, we demonstrate that MAT can effectively improve model robustness and outperform state-of-the-art methods under white-box attacks. In addition, we show that MAT can also mitigate the robustness trade-off among different perturbation types. Specially, we train specialist models that learn to defend a specific perturbation type and a generalist model that learns to defend multiple perturbation types by learning from the specialists, which brings as much as 13.4% accuracy gain to AT baselines against the union of$l_{\infty} $ ,$l_{2}$ , and$l_{1}$ attacks. Our results show the effectiveness of the proposed method and demonstrate that collaborative learning is an effective strategy for designing robust models.- 17 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 17 (15566013) , 2364-2377
Mutual Adversarial Training: Learning Together is Better Than Going Alone
Recent studies have shown that robustness to adversarial attacks can be transferred across deep neural networks. In other words, we can make a weak model more robust with the help of a strong teacher model. In this paper, we ask if models can “learn together” and “teach each other” to achieve better robustness instead of learning from a static teacher. We study how interactions among models enhance robustness via knowledge distillation. We propose mutual adversarial training (MAT), in which multiple models are trained together and share the knowledge of adversarial examples to achieve improved robustness. MAT allows robust models to explore a larger space of adversarial samples and find more robust feature spaces and decision boundaries. Through extensive experiments on the CIFAR-10, CIFAR-100, and mini-ImageNet datasets, we demonstrate that MAT can effectively improve model robustness and outperform state-of-the-art methods under white-box attacks. In addition, we show that MAT can also mitigate the robustness trade-off among different perturbation types. Specially, we train specialist models that learn to defend a specific perturbation type and a generalist model that learns to defend multiple perturbation types by learning from the specialists, which brings as much as 13.4% accuracy gain to AT baselines against the union of$l_{\infty} $ ,$l_{2}$ , and$l_{1}$ attacks. Our results show the effectiveness of the proposed method and demonstrate that collaborative learning is an effective strategy for designing robust models.- 9 February 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 3 (2) , 240-251
ATFaceGAN: Single Face Semantic Aware Image Restoration and Recognition From Atmospheric Turbulence
Degradation due to atmospheric turbulence is common while capturing images at long ranges. To mitigate the degradation due to turbulence which includes deformation and blur, we propose a generative single frame restoration algorithm which disentangles the blur and deformation due to turbulence and reconstructs a restored image. The disentanglement is achieved by decomposing the distortion due to turbulence into blur and deformation components using a deblur generator and a deformation correction generator, respectively. Two paths of restoration are implemented to regularize the disentanglement and generate two restored images from a given degraded image. A fusion function combines the features of the restored images to reconstruct a sharp image with rich details. Adversarial and perceptual losses are added to reconstruct a sharp image and suppress the artifacts respectively. Extensive experiments demonstrate the effectiveness of the proposed restoration algorithm, which achieves satisfactory performance for face restoration and face recognition tasks.
- 9 February 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 3 (2) , 240-251
ATFaceGAN: Single Face Semantic Aware Image Restoration and Recognition From Atmospheric Turbulence
Degradation due to atmospheric turbulence is common while capturing images at long ranges. To mitigate the degradation due to turbulence which includes deformation and blur, we propose a generative single frame restoration algorithm which disentangles the blur and deformation due to turbulence and reconstructs a restored image. The disentanglement is achieved by decomposing the distortion due to turbulence into blur and deformation components using a deblur generator and a deformation correction generator, respectively. Two paths of restoration are implemented to regularize the disentanglement and generate two restored images from a given degraded image. A fusion function combines the features of the restored images to reconstruct a sharp image with rich details. Adversarial and perceptual losses are added to reconstruct a sharp image and suppress the artifacts respectively. Extensive experiments demonstrate the effectiveness of the proposed restoration algorithm, which achieves satisfactory performance for face restoration and face recognition tasks.