Author name cluster

Dan Ma

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

3 papers

1 author row

AAAI Conference 2026 Conference Paper

Beyond Fully Supervised Pixel Annotations: Scribble-Driven Weakly-Supervised Framework for Image Manipulation Localization

Songlin Li
Guofeng Yu
Zhiqing Guo
Yunfeng Diao
Dan Ma
Gaobo Yang

Deep learning-based image manipulation localization (IML) methods have achieved remarkable performance in recent years, but typically rely on large-scale pixel-level annotated datasets. To address the challenge of acquiring high-quality annotations, some recent weakly supervised methods utilize image-level labels to segment manipulated regions. However, the performance is still limited due to insufficient supervision signals. In this study, we explore a form of weak supervision that improves the annotation efficiency and detection performance, namely scribble annotation supervision. We re-annotated mainstream IML datasets with scribble labels and propose the first scribble-based IML (Sc-IML) dataset. Additionally, we propose the first scribble-based weakly supervised IML framework. Specifically, we employ self-supervised training with a structural consistency loss to encourage the model to produce consistent predictions under multi-scale and augmented inputs. In addition, we propose a prior-aware feature modulation module (PFMM) that adaptively integrates prior information from both manipulated and authentic regions for dynamic feature adjustment, further enhancing feature discriminability and prediction consistency in complex scenes. We also propose a gated adaptive fusion module (GAFM) that utilizes gating mechanisms to regulate information flow during feature fusion, guiding the model toward emphasizing potential tampered regions. Finally, we propose a confidence-aware entropy minimization loss. This loss dynamically regularizes predictions in weakly annotated or unlabeled regions based on model uncertainty, effectively suppressing unreliable predictions. Experimental results show that our method outperforms existing fully supervised approaches in terms of average performance both in-distribution and out-of-distribution.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Uncovering and Mitigating Destructive Multi-Embedding Attacks in Deepfake Proactive Forensics

Lixin Jia
Haiyang Sun
Zhiqing Guo
Yunfeng Diao
Dan Ma
Gaobo Yang

With the rapid evolution of deepfake technologies and the wide dissemination of digital media, personal privacy is facing increasingly serious security threats. Deepfake proactive forensics, which involves embedding imperceptible watermarks to enable reliable source tracking, serves as a crucial defense against these threats. Although existing methods show strong forensic ability, they rely on an idealized assumption of single watermark embedding, which proves impractical in real-world scenarios. In this paper, we formally define and demonstrate the existence of Multi-Embedding Attacks (MEA) for the first time. When a previously protected image undergoes additional rounds of watermark embedding, the original forensic watermark can be destroyed or removed, rendering the entire proactive forensic mechanism ineffective. To address this vulnerability, we propose a general training paradigm named Adversarial Interference Simulation (AIS). Rather than modifying the network architecture, AIS explicitly simulates MEA scenarios during fine-tuning and introduces a resilience-driven loss function to enforce the learning of sparse and stable watermark representations. Our method enables the model to maintain the ability to extract the original watermark correctly even after a second embedding. Extensive experiments demonstrate that our plug-and-play AIS training paradigm significantly enhances the robustness of various existing methods against MEA.

PDF Details DOI

JBHI Journal 2024 Journal Article

Automated Prediction of Infant Cognitive Development Risk by Video: A Pilot Study

Shengjie Ji
Dan Ma
Lunxin Pan
Wenan Wang
Xiaohang Peng
Joan Toluwani Amos
Honorine Niyigena Ingabire
Min Li

Objective: Cognition is an essential human function, and its development in infancy is crucial. Traditionally, pediatricians used clinical observation or medical imaging to assess infants’ current cognitive development (CD) status. The object of pediatricians’ greater concern is however their future outcomes, because high-risk infants can be identified early in life for intervention. However, this opportunity has not yet been realized. Fortunately, some recent studies have shown that the general movement (GM) performance of infants around 3–4 months after birth might reflect their future CD status, which gives us an opportunity to achieve this goal by cameras and artificial intelligence. Methods: First, infants’ GM videos were recorded by cameras, from which a series of features reflecting their bilateral movement symmetry (BMS) were extracted. Then, after at least eight months of natural growth, the infants’ CD status was evaluated by the Bayley Infant Development Scale, and they were divided into high-risk and low-risk groups. Finally, the BMS features extracted from the early recorded GM videos were fed into the classifiers, using late infant CD risk assessment as the prediction target. Results: The area under the curve, recall and precision values reached 0. 830, 0. 832, and 0. 823 for two-group classification, respectively. Conclusion: This pilot study demonstrates that it is possible to automatically predict the CD of infants around the age of one year based on their GMs recorded early in life. Significance: This study not only helps clinicians better understand infant CD mechanisms, but also provides an economical, portable and non-invasive way to screen infants at high-risk early to facilitate their recovery.

Details DOI