The pervasive spread of fake news, particularly through manipulated images, presents a consequential negative impact on society. To prevent fake news images from misleading the public, existing methods focus on verifying the authenticity of news images but ignore source traceability, leaving a gap in creating a complete forensic chain for reliable fake news detection. To simultaneously achieve the goals of authenticity verification and source tracing, we propose a proactive image tagging approach based on a design of Disentangled Invertible Neural Networks (DINN). It can simultaneously embed the dual-tags, i.e., authenticable tag and traceable tag, into each news image prior to publication, allowing for separate extraction for authenticity verification and source tracing. Within the proposed DINN, we design a parallel Feature Aware Projection Module (FAPM) to assist DINN in preserving essential tag information, thereby improving extraction accuracy. In addition, we introduce a Distance Metric-Guided Module (DMGM) that learns asymmetric one-class representations, enabling the dual-tags to exhibit different robustness performances under malicious manipulations. Extensive experiments on diverse datasets and unseen manipulations demonstrate that the proposed tagging approach achieves promising performances on both authenticity verification and source tracing for reliable fake news detection and outperforms the prior works.

With the rapid development of e-health technology and artificial intelligence, healthcare services have undergone significant digitalization and intelligence transformation. This drives the emergence of innovative models such as personalized care and telemedicine, making electronic medical information sharing a critical research focus. However, due to the high sensitivity of medical data, existing sharing platforms face challenges, including insufficient privacy protection, inadequate guarantees of data integrity and authenticity, and reliance on centralized systems. These issues can lead to risks, including identity information leakage, data tampering, and single points of failure. To address these challenges, this article proposes a blockchain-based framework for electronic medical information sharing, namely, PCTRS, which stands for privacy and control in electronic medical records (EMRs) through blockchain and threshold traceable ring signatures. The framework integrates InterPlanetary File System (IPFS) for off-chain storage, reducing costs while enhancing data security through elliptic curve Diffie-Hellman (ECDHE) ephemeral key exchange and Advanced Encryption Standard Galois/Counter Mode (AES-GCM) encryption. By integrating smart contracts with a threshold-based traceable ring signature scheme, the system achieves an optimal balance between privacy protection and transparency. Furthermore, the system implements a reputation mechanism combined with decentralized arbitration to ensure reliability, effectively mitigating security risks in medical data sharing.

Current deepfake detectors rely on spatial-frequency analysis for low-level artifacts or CLIP for semantic priors. However, both suffer from representation incompleteness, lacking either semantic depth or fine-grained frequency details. Furthermore, they share a critical vulnerability to spurious correlations, which severely hinders cross-domain generalization. To tackle these issues, we propose ReCLIP-SFD, a synergistic framework designed for highly generalizable deepfake detection. To overcome representation incompleteness, we design an asymmetric dual-branch architecture that explicitly integrates a CLIP-based spatial branch with an SRM-guided CNN frequency branch. This structure achieves adaptive synergy between high-level semantics and subtle low-level forensic artifacts via a Residual-Guided Correction Module (RGCM). Furthermore, to mitigate spurious correlations, we introduce the Teacher-Student Debiasing Framework (TSDF), which employs cross-view imitability to suppress dataset-specific shortcuts. TSDF encourages bidirectional prediction alignment between a Teacher and a Student under independent stochastic augmentations, guiding the detector toward transferable forgery evidence rather than dataset-specific spurious correlations. Trained on a single dataset, ReCLIP-SFD achieves state-of-the-art performance, with average Area Under the Curve (AUC) scores of 90.6% and 94.2% at the frame and video levels, respectively, on cross-dataset benchmarks. Furthermore, it demonstrates superior generalization across extensive cross-manipulation and AI-Generated Content (AIGC) detection evaluations. Our work underscores the necessity of jointly addressing representation incompleteness and spurious correlations to build truly generalizable deepfake detectors.

While the widespread adoption of diffusion models in image generation has showcased remarkable capabilities, it has also inadvertently opened the door to malicious exploitation. Recent research has primarily concentrated on protecting images from the misuse of diffusion-based customized generation (CG). However, these approaches often overlook that image details can still be enhanced through diffusion-based super-resolution (SR) techniques, significantly increasing the risks of personal image leakage and abuse. To combat these multifaceted risks, we propose the Zero Matrix-guided Adaptive Image Vaccine (ZMAIV) framework. Specifically, we introduce the Self-attention Removal strategy, tailored for CG, which disrupts the model's core mechanism of focusing on sensitive spaces. Concurrently, the High-frequency Removal strategy is proposed to impede the high-frequency details reconstruction of SR. These defense strategies effectively dismantle the underlying mechanisms that facilitate unauthorized data extrapolation. Moreover, the proposed Adaptive Space Search Attack precisely targets critical spaces within images for vaccine injection, optimizing perturbation placement to minimize perturbation conflict while maintaining defense performance. Extensive experiments demonstrate that the proposed ZMAIV outperforms the state-of-the-arts in the aspects of simultaneously defending against diffusion-based CG and SR, affirming its superiority in safeguarding visual content against these dual threats.

Object detection has become crucial due to its extensive applications across various industries. However, the data used to train these models is often sensitive and proprietary, raising significant concerns about its security and unauthorized usage. Unlearnable examples (UEs) represent a promising strategy to safeguard proprietary datasets by embedding imperceptible perturbations that degrade model performance when such data is used during training. This paper explores UEs tailored specifically for object detection tasks, which pose unique challenges due to the multi-task nature of object detection. We propose a novel framework that generates robust and effective UEs, and significantly degrades object detector performance while maintaining imperceptibility. Comprehensive experiments demonstrate the resilience of the proposed UEs against various countermeasures, underscoring their potential as a practical solution for protecting data in object detection.

Deep Image Hiding (DIH) aims to imperceptibly hide images within image. To improve its security performance, some DIH methods design Security Metrics (SMs) to guide the learning of their hiding networks. However, these methods focus on optimizing their anti-steganalysis ability on specific SMs, resulting in inferior generalization ability. To overcome these limitations, in this paper, we introduce meta-learning into DIH and propose Meta Security Metric-based DIH (MSM-DIH). In the MSM-DIH, the Invertible Neural Network (INN)-based hiding network is learned under the guidance of a learnable meta SM generalized from multiple fixed source SMs, and each SM is composed of a metric network and a contrastive loss function. Specifically, MSM-DIH is trained with bi-level optimization. In the outer optimization, a meta SM is learned to assign higher security scores for more advanced stego images. Besides, the domain knowledge of steganalysis is transferred from the multiple pre-trained source metric networks to the meta metric network, so as to enhance the generalization ability of the meta SM. In the inner optimization, the hiding network is learned to generate more secure stego images according to the learned meta SM. Experimental results show that our MSM-DIH has achieved the best security performance in most cases.

Despite the advancements that deep learning has brought to medical image analysis (MIA), protecting the privacy of images remains a challenge. In a client-server MIA framework, especially after deployment, patients' private medical images can be easily captured by attackers from the transmission channel or malicious third-party servers. Previous MIA privacy-enhancing methods, whether based on distortion or homomorphic encryption, expose the fact that the transmitted images are medical images or transform the images into semantic-lacking noise. This tends to alert attackers, thereby falling into a cat-and-mouse game of theft and protection. To address this issue, we propose a covert MIA framework based on deep image hiding, namely HideMIA, which secures medical images by embedding them within natural cover images that are unlikely to raise suspicion. By directly analyzing the hidden medical images in the steganographic domain, HideMIA makes it difficult for attackers to notice the presence of medical images. Specifically, we propose the Mixture-of-Difference-Convolutions (MoDC) and Asymmetric Wavelet Attention (AsyWA) to enable HideMIA to conduct fine-grained analysis on each wavelet sub-band within the steganographic domain, mining features that are specific to medical images. Moreover, to reduce resource consumption on client devices, we design function-aligned knowledge distillation to obtain a lightweight hiding network, namely LightIH. Extensive experiments on six medical datasets demonstrate that our HideMIA achieves superior MIA performance and protective imperceptibility on medical image segmentation and classification.

Ensuring data privacy and protection has become paramount in the era of deep learning. Unlearnable examples are proposed to mislead the deep learning models and prevent data from unauthorized exploration by adding small perturbations to data. However, such perturbations (e.g., noise, texture, color change) predominantly impact low-level features, making them vulnerable to common countermeasures. In contrast, semantic images with intricate shapes have a wealth of high-level features, making them more resilient to countermeasures and potential for producing robust unlearnable examples. In this paper, we propose a Deep Hiding (DH) scheme that adaptively hides semantic images enriched with high-level features. We employ an Invertible Neural Network (INN) to invisibly integrate predefined images, inherently hiding them with deceptive perturbations. To enhance data unlearnability, we introduce a Latent Feature Concentration module, designed to work with the INN, regularizing the intra-class variance of these perturbations. To further boost the robustness of unlearnable examples, we design a Semantic Images Generation module that produces hidden semantic images. By utilizing similar semantic information, this module generates similar semantic images for samples within the same classes, thereby enlarging the inter-class distance and narrowing the intra-class distance. Extensive experiments on CIFAR-10, CIFAR-100, and an ImageNet subset, against 18 countermeasures, reveal that our proposed method exhibits outstanding robustness for unlearnable examples, demonstrating its efficacy in preventing unauthorized data exploitation.

Ensuring the legal usage of deep models is crucial to promoting trustable, accountable, and responsible artificial intelligence innovation. Current passport-based methods that obfuscate model functionality for license-to-use and ownership verifications suffer from capacity and quality constraints, as they require retraining the owner model for new users. They are also vulnerable to advanced Expanded Residual Block ambiguity attacks. We propose Steganographic Passport, which uses an invertible steganographic network to decouple license-to-use from ownership verification by hiding the user's identity images into the owner-side passport and recovering them from their respective user-side passports. An irreversible and collision-resistant hash function is used to avoid exposing the owner-side passport from the derived user-side passports and increase the uniqueness of the model signature. To safeguard both the passport and model's weights against advanced ambiguity attacks, an activation-level obfuscation is proposed for the verification branch of the owner's model. By jointly training the verification and deployment branches, their weights become tightly coupled. The proposed method supports agile licensing of deep models by providing a strong ownership proof and license accountability without requiring a separate model retraining for the admission of every new user. Experiment results show that our Steganographic Passport outperforms other passport-based deep model protection methods in robustness against various known attacks.

Taking the steganalytic discriminators as the adversaries, the existing Generative Adversarial Networks (GAN)-based steganographic approaches learn the implicit cost functions to measure the embedding distortion for steganography. However, the steganalytic discriminators in these approaches are trained by the stego-samples with insufficient diversity, and their network structures offer very limited representational capacity. As a result, these steganalytic discriminators will not exhibit robustness to various steganographic patterns, which causes learning suboptimal cost functions, thus compromising the anti-steganalysis capability. To address this issue, we propose a novel GAN-based steganographic approach, in which the Diversified Inverse-Adversarial Training (DIAT) strategy and the Steganalytic Feature Attention (SteFA) structure are designed to train a robust steganalytic discriminator. Specifically, the DIAT strategy provides the steganalytic discriminator with an expanded feature space by generating diversified adversarial stego-samples; the SteFA structure enables the steganalytic discriminator to capture more various steganalytic features by employing the channel-attention mechanism on higher-order statistics. Consequently, the steganalytic discriminator can build a more precise decision boundary to make it more robust, which facilitates learning a superior steganographic cost function. Extensive experiments demonstrate that the proposed steganographic approach achieves promising anti-steganalysis capability over the state-of-the-arts under the same embedding payloads.

Ensuring data privacy and protection has become paramount in the era of deep learning. Unlearnable examples are proposed to mislead the deep learning models and prevent data from unauthorized exploration by adding small perturbations to data. However, such perturbations (e.g., noise, texture, color change) predominantly impact low-level features, making them vulnerable to countermeasures like adversarial training, data augmentations, and preprocessing. In contrast, semantic images with intricate shapes have a wealth of high-level features, making them more resilient to countermeasures and potential for producing robust unlearnable examples. In this paper, we propose a Deep Hiding (DH) scheme that adaptively hides semantic images enriched with high-level features. We employ an Invertible Neural Network (INN) to invisibly integrate predefined images, inherently hiding them with deceptive perturbations. To enhance data unlearnability, we introduce a Latent Feature Concentration module, designed to work with the INN, regularizing the intra-class variance of these perturbations. To further boost the robustness of unlearnable examples, we design a Semantic Images Generation module that produces hidden semantic images. By utilizing similar semantic information, this module generates similar semantic images for samples within the same classes, thereby enlarging the inter-class distance and narrowing the intra-class distance. Extensive experiments on CIFAR-10, CIFAR-100, and ImageNet-subset, against 12 countermeasures, reveal that our proposed method exhibits state-of-the-art ro- bustness for unlearnable examples, demonstrating its efficacy in data protection.

The data in the blockchain cannot be tampered with and the users are anonymous, which enables the blockchain to be a natural carrier for covert communication. However, the existing methods of covert communication in blockchain suffer from the predefined channel structure, the capacity of a single transaction is not high, and the fixed transaction behaviors will lower the concealment of the communication channel. Therefore, this paper proposes a derivation matrix-based covert communication method in blockchain. It uses dual-key to derive two types of blockchain addresses and then constructs an address matrix by dividing addresses into multiple layers to make full use of the redundancy of addresses. Subsequently, to solve the problem of the lack of concealment caused by the fixed transaction behaviors, divide the rectangular matrix into square blocks with overlapping regions and then encrypt different blocks sequentially to make the transaction behaviors of the channel addresses match better with those of the real addresses. Further, the linear congruence algorithm is used to generate random sequence, which provides a random order for blocks encryption, and thus enhances the security of the encryption algorithm. Experimental results show that this method can effectively reduce the abnormal transaction behaviors of addresses while ensuring the channel transmission efficiency.

Robust watermarking requires finding invariant features under multiple attacks to ensure correct extraction. Deep learning has extremely powerful in extracting features, and watermarking algorithms based on deep learning have attracted widespread attention. Most existing methods use 3 x 3 small kernel convolution to extract image features and embed the watermarking. However, the effective perception fields for small kernel con-volution are extremely confined, so the pixels that each watermarking can affect are restricted, thus limiting the performance of the watermarking. To address these problems, we propose a watermarking network based on large kernel convolution and adaptive weight assignment for loss functions. It uses large-kernel depth-wise convolution to extract features for learning large-scale image information and subsequently projects the watermarking into a high -dimensional space by 1 x 1 convolution to achieve adaptability in the channel dimension. Subsequently, the modification of the embedded watermarking on the cover image is extended to more pixels. Because the magnitude and convergence rates of each loss function are different, an adaptive loss weight assignment strategy is proposed to make the weights participate in the network training together and adjust the weight dynamically. Further, a high-frequency wavelet loss is proposed, by which the watermarking is restricted to only the low-frequency wavelet sub-bands, thereby enhancing the robustness of watermarking against image compression. The experimental results show that the peak signal-to-noise ratio (PSNR) of the encoded image reaches 40.12, the structural similarity (SSIM) reaches 0.9721, and the watermarking has good robustness against various types of noise.

As a promising technique of resisting steganalysis detection, generative steganography usually generates a new image driven by secret information as the stego-image. However, it generally encodes secret information as entangled features in a non-distribution-preserving manner for the stego-image generation, which leads to two common issues: 1) limited accuracy of information extraction, and 2) low security in the feature-domain. To address the above limitations, we propose a generative steganographic framework via auto-generation of semantic object contours, in which a given secret message is encoded as the disentangled features, i.e., object-contours, in a distribution-preserving manner for the stego-image generation. In this framework, we propose a contour generative adversarial nets (CtrGAN) consisting of a contour-generator and a contour-discriminator, which are adversarially trained with reinforcement learning. To realize the generative steganography, by using the contour-generator of the trained CtrGAN, a contour point selection (CPS)-based encoding strategy is designed to encode the secret message as the contours. Then, the BicycleGAN is employed to transform the generated contours to the corresponding stego-image. Extensive experiments demonstrate that the proposed steganographic approach outperforms the state-of-the-arts in terms of information extraction accuracy, especially under common image attacks, and feature-domain security.

针对当前指纹识别系统容易遭受伪造指纹欺骗攻击的问题,提出一种基于纹理特征融合的指纹活性检测算法.通过设计边缘纹理增强(ETE)和对称差分统计(SDS)2 种脊线纹理特征描述算子来表示真假指纹的显著性纹理,前者用来提取指纹图像脊线的方向纹理信息,后者用来描述邻域内脊线的频率纹理信息.首先,利用感兴趣区域(ROI)提取算法对指纹图像进行预处理,以消除指纹图像中背景空白噪声的干扰;然后,利用ETE和SDS分别提取指纹的脊线纹理特征;接着,统计上述2 类特征的直方图,描述真假指纹的纹理特征;最后,将生成的特征输入支持向量机(SVM)中进行训练和测试.在 LiveDet 2011 指纹数据集的测试中,分别使用 Biometrika、Italdata、Sagem 3 种传感器,且与Best、韦伯局部描述算子(WLD)、局部相位量化(LPQ)和局部二值模式(LBP)4 种指纹检测算法进行了比较,该文算法的检测性能优于其余方法,能够完成当前的活性检测任务.LiveDet 2013 数据集使用Biometrika、Italdata和Swipe 3 种传感器,通过与WLD、不变梯度直方图(HIG)、统一局部二值模式(ULBP)、深度表征结构优化(DRAO)和Winner 5 种指纹活性检测方法对比,该文算法的指纹活性检测准确率有一定的提升.

To detect the existing steganographic algorithms, recent steganalysis methods usually train a Convolutional Neural Network (CNN) model on the dataset consisting of corresponding paired cover/stego-images. However, it is inefficient and impractical for those steganalysis tools to completely retrain the CNN model to make it effective against both the existing steganographic algorithms and a new emerging steganographic algorithm. Thus, existing steganalysis models usually lack dynamic extensibility for new steganographic algorithms, which limits their application in real-world scenarios. To address this issue, we propose an accurate parameter importance estimation (APIE) based-continual learning scheme for steganalysis. In this scheme, when a steganalysis model is trained on the new image dataset generated by the new steganographic algorithm, its network parameters are effectively and efficiently updated with sufficient consideration of their importance evaluated in the previous training process. This approach can guide the steganalysis model to learn the patterns of the new steganographic algorithm without significantly degrading the detectability against the previous steganographic algorithms. Experimental results demonstrate the proposed scheme has promising extensibility for new emerging steganographic algorithms.

To prevent fake news images from misleading the public, it is desirable not only to verify the authenticity of news images but also to trace the source of fake news, so as to provide a complete forensic chain for reliable fake news detection. To simultaneously achieve the goals of authenticity verification and source tracing, we propose a traceable and authenticable image tagging approach that is based on a design of Decoupled Invertible Neural Network (DINN). The designed DINN can simultaneously embed the dual-tags, \textit{i.e.}, authenticable tag and traceable tag, into each news image before publishing, and then separately extract them for authenticity verification and source tracing. Moreover, to improve the accuracy of dual-tags extraction, we design a parallel Feature Aware Projection Model (FAPM) to help the DINN preserve essential tag information. In addition, we define a Distance Metric-Guided Module (DMGM) that learns asymmetric one-class representations to enable the dual-tags to achieve different robustness performances under malicious manipulations. Extensive experiments, on diverse datasets and unseen manipulations, demonstrate that the proposed tagging approach achieves excellent performance in the aspects of both authenticity verification and source tracing for reliable fake news detection and outperforms the prior works.

To realize secure communication, steganography is usually implemented by embedding secret information into an image selected from a natural image dataset, in which the fractal images have occupied a considerable proportion. To detect those stego-images generated by existing steganographic algorithms, recent steganalysis models usually train a Convolutional Neural Network (CNN) on the dataset consisting of paired cover/stego-images. However, it is inefficient and impractical for those steganalysis models to completely retrain the CNN model to make it effective for detecting a new emerging steganographic algorithm while maintaining the ability to detect the existing steganographic algorithms. Thus, those steganalysis models usually lack dynamic extensibility for new steganographic algorithms, which limits their application in real-world scenarios. To address this issue, we propose an accurate parameter importance estimation (APIE)-based continual learning scheme for steganalysis. In this scheme, when a steganalysis model is trained on a new image dataset generated by a new emerging steganographic algorithm, its network parameters are effectively and efficiently updated with sufficient consideration of their importance evaluated in the previous training process. This scheme can guide the steganalysis model to learn the patterns of the new steganographic algorithm without significantly degrading the detectability against the previous steganographic algorithms. Experimental results demonstrate the proposed scheme has promising extensibility for new emerging steganographic algorithms.

Steganography is an essential way to ensure secure communication in networks. Most steganographic algorithms imperceptibly embed secret information into an existing cover image. However, they generally cannot find a good trade-off between embedding capacity and security, as the existing covers available for users are usually far from optimal for embedding. To address this issue, instead of directly using the existing cover images, we propose a cover enhancement scheme for high-capacity image steganography, in which textured objects are generated and adaptively pasted to an existing cover based on the estimated embedding probability maps. Specifically, by estimating the embedding probability map of the cover image, we locate the high-embedding-cost region (HECR), which is inappropriate for embedding. Then, a textured object is generated by the conditional generative adversarial networks with the input of an affinely transformed object mask, and then is pasted to the located HECR for steganography. Since the image regions inappropriate for embedding are replaced by the textured object regions, the proposed scheme can provide a much higher embedding capacity for the state-of-the-art steganographic approaches. Extensive experiments demonstrate that the proposed scheme provides high embedding capacity, i.e., about 2.5 times higher than the state-of-the-art steganographic methods, and comparable anti-detectability to those methods.

Graphical password systems have received significant attention as one potential solution to the need for more usable authentication, but graphical passwords are often considered prone to shoulder-surfing attacks. In this paper, in order to build a balance between usability and security for authentication, we propose a new graphical password scheme by combining a specific verification grid with map slipping strategy. In the proposed scheme, a set of fixed grids are pregenerated on a map. During the registration process, the user is not only asked to select several points in sequence on the map to form a password route, but also required to choose one of the pregenerated grids as a specific verification grid for the subsequent authentication. The password route and the specific verification grid together form the complete graphical password. During the authentication process, the user needs to slip the map to let each point on the password route sequentially fit inside the specific verification grid which is already remembered by the user but difficult to be detected by the attackers. With the specific verification grid and the map slipping strategy, the proposed scheme can effectively defend against shoulder-surfing attacks. Meanwhile, the password points are represented as coordinates on the map; therefore, the proposed scheme has a negligible storage burden. The comparative experiments show that, using the proposed scheme, the success rate of shoulder-surfing defense can be increased by 37% to 56% with different grid sizes and password point numbers, and the usability of passwords can also be improved by 3% to 6%. Therefore, the proposed scheme can achieve good shoulder-surfing defense and reasonable usability simultaneously.