Extracting discriminative forensic artifacts from high-dimensional latent spaces is pivotal for Image Manipulation Localization (IML). Nevertheless, the intrinsic feature entanglement arising from subtle structural discrepancies continues to impede precise localization. Specifically, manipulation artifacts are frequently overwhelmed by coherent background textures or ’soft boundaries’, rendering the trace-rich features inextricably mixed with intrinsic image content. To disentangle these intertwined representations, we propose the Decoupled Contrastive Learning Network (DCL-Net) for robust image tampering localization. Leveraging a Vision Transformer (ViT) backbone integrated with a Simple Feature Pyramid Network (SFPN), DCL-Net incorporates a novel Feature Decoupling Module (FDM). The FDM explicitly disentangles the feature space into foreground, background, and uncertainty regions-thereby effectively isolating manipulation cues from coherent background textures and capturing the transitional nature of ambiguous boundaries. Furthermore, to align the optimization objective with the intrinsic structure of manipulation traces, we introduce a prior-guided contrastive learning strategy that explicitly ’pushes away’ manipulated features from authentic and uncertain components. Finally, the Contrast Feature Aggregation Module (CFAM) employs a two-stage attention mechanism to refine these disentangled features, further suppressing redundant background details. Extensive experiments across five public benchmarks demonstrate that DCL-Net delivers state-of-the-art localization performance and exhibits strong robustness against common distortions.

Coverless image steganography has attracted considerable attention due to its ability to evade traditional steganalysis techniques. However, many existing coverless methods face difficulties in reliably recovering secret information when subjected to geometric attacks. To overcome this limitation, we propose a novel approach based on deep unsupervised hashing for coverless steganography, designed to improve resilience against geometric attacks. The method begins by developing a robust model for hash sequence generation. To further strengthen the model's resistance to a variety of attacks, we integrate an attention mechanism using depth- wise separable convolutions, combined with a Transformer module that captures global context and long-range dependencies. In this framework, the sender processes an image through the trained model to generate a hash sequence, while the receiver uses the same model to recover the original secret information from the stego image, ensuring the security of the steganographic process. Additionally, we employ an inverted index to facilitate efficient and rapid matching of stego images. Experimental evaluations show that the proposed method outperforms existing approaches in terms of robustness to both geometric and non-geometric attacks. Specifically, as the intensity of geometric attacks increases, our method consistently achieves high information recovery performance, with an average robustness score of 93.6% across several datasets. These results highlight the enhanced robustness and security of the proposed coverless steganographic approach.

The objective of radiology report generation is to alleviate the burden on physicians in drafting reports, thereby improving generation efficiency and reducing patient waiting times. In recent years, there has been a growing emphasis on imaging-based monitoring technologies within healthcare, with a particular focus on the precise detection and interpretation of subtle changes. Likewise, as individual X-rays often exhibit minimal differences, pathologies are frequently concealed within intricate details, making accurate report generation a challenge. To address this issue, we propose a novel method consisting of three key modules: the Self-Attention Mamba Module (Self-Mamba), the Cross-Attention Mamba Module (Cross-Mamba), and the Sparse Mask Loss Function (Sparse-Loss). When processing an X-ray, we use a similar approach to human observation, focusing first on the overall structure and then focusing on identifying possible focal areas. For this purpose, we design Self-Mamba module to extract the features of abnormal areas in X-ray images through global information modeling. The proposed Cross-Mamba module enhances the consistency of medical images and radiology reports by optimizing the ability of cross-modal interaction between the two. Sparse-Loss function is proposed to alleviate the problem of unbalance of positive and negative samples by taking advantage of its sparsity. Experimental results show that our approach outperforms existing models on several metrics and achieves excellent performance on two publicly available datasets, IU-Xray and COV-CTR.

Cross-modal retrieval techniques, with their efficient search capabilities, have garnered considerable attention in both industry and academia. Among them, image-text cross-modal retrieval, a classic task in the cross-modal domain, has been widely applied in internet applications. However, image data in cross-modal retrieval applications often contains rich personal sensitive information, which is typically stored in plaintext format, posing a risk of privacy leakage. Due to the unique semantic and heterogeneous gap in the cross-modal space, existing single-modality image privacy protection techniques cannot be directly applied to cross-modal scenarios. To address this issue, this paper proposes a privacy-preserving cross-modal retrieval method based on disentangled soft-label alignment(PPDSA). Firstly, to protect the privacy of image data, thumbnail preserving encryption is used in this paper. This method aims to encrypt the image while ensuring the feasibility of graph-text cross-modal feature representation learning, so as to realize cross-modal encrypted retrieval. Second, to further improve the retrieval effect of encrypted images and normal text, this paper introduces a disentangled soft label alignment technique. Specifically, a teacher model is used to obtain the soft label matrix, and cross-modal (encrypted image-text pair) and single-modal (encrypted image pair) soft label alignment techniques are designed to capture more fine-grained semantic representation information and reduce the interference of false positives on similarity recognition. A series of experiments on Flickr30k and MSCOCO datasets show that the proposed method not only effectively protects image privacy in cross-modal retrieval, but also improves the retrieval performance by 6.7% and 3.7% respectively compared with the CLIP baseline.

Abstract Information mapping is a widely adopted strategy in information hiding, leveraging the inherent features of carriers to convey hidden information without altering the carrier itself, thus maintaining integrity and avoiding detection. However, current mapping-based techniques face significant challenges, including potential data loss during transmission, limited capacity of carriers like images, and reliance on costly third-party storage solutions. To address these limitations, we introduce BlkInfoM, an innovative algorithm that utilizes blockchain’s decentralized, immutable, and traceable properties as a novel data source for secure information hiding. BlkInfoM leverages blockchain transaction data, such as Merkle hash values, timestamps, and locations, in combination with a reversible ASCII-based binary encoding to enable precise information-to-block matching. Experimental results demonstrate that BlkInfoM not only improves the success rate and efficiency of information mapping compared to traditional methods but also reduces operational costs by eliminating the need for third-party storage. This work highlights the potential of blockchain technology to revolutionize information hiding, offering enhanced security, scalability, and cost-effectiveness.

Screen-shooting watermarking technology plays a critical role in copyright protection and traceability. However, existing methods often lack sufficient robustness under strong noise interference and tend to introduce noticeable visual artifacts when embedding watermarks in smooth image regions, thereby degrading visual quality and increasing the risk of watermark exposure. To address these limitations, this paper proposes a Contrastive Learning and Mask-guided Embedding (CLME) framework for robust screen-shooting watermarking. The framework comprises two key components: (1) a mask-guided watermark embedding module that utilizes a Residual Dense Feature Extraction Block (RDFEB) and an Attention Mask Generation Block (AMGB) to adaptively embed watermarks into texture-rich regions, improving watermark invisibility; and (2) a contrastive learning-based watermark decoding network that employs contrastive loss to enhance the consistency of decoded features by treating features from the same watermarked image under different noise conditions as positive samples and features from different watermarked images as negative samples, thereby improving the robustness of watermark extraction. Experimental results demonstrate that the proposed CLME framework outperforms existing methods in terms of both robustness and visual quality. Specifically, at a shooting distance of 100 cm and a shooting angle of 40°, the watermark extraction accuracy reaches 99.58%, and the peak signal-to-noise ratio (PSNR) of the watermarked images reaches 42.624 dB, highlighting the framework’s strong potential for real-world applications.

Deep cross-modal retrieval, with its effective and efficient search capabilities, has gained widespread adoption in today’s media-sharing practices yet raises concerns regarding potential threats to user data privacy. The cutting-edge data-centric countermeasures usually adopt adversarial learning, i.e., laboriously crafting the proper perturbation for each image, resulting in the noticeable noise in adversarial examples that greatly undermines the aesthetic appeal of image sharing. To address this issue, we propose a novel Model-centric Cross-modal Privacy-preserving framework (MCP), wherein the pre-defined invisible backdoor is seamlessly integrated into the global retrieval model via backdoor learning, thereby effectively preventing shared images containing such triggers from being retrieved. Specifically, we introduce a simple yet effective cross-modal backdoor learning algorithm that alternately optimizes two losses: 1) a privacy-preserving loss for perturbing retrieval with a user-injected trigger and 2) the standard utility loss for maintaining normal retrieval performance. Compared to state-of-the-art methods, MCP excels in providing excellent stealthiness, manifesting in a notable improvement of approximately 100% in SSIM metrics. Furthermore, it achieves an outstanding privacy-preserving (backdoor) success rate, as evidenced by a substantial mAP reduction of 22.3% (for FashionVC), 11.5% (for NUS-WIDE), and 21.8% (for MIRFlickr-25K) in poisoned retrieval, while maintaining similar normal retrieval performance. Additionally, MCP exhibits robust resistance against potential black-box defenses (e.g., trigger filtering) and white-box defenses (e.g., fine-tuning and model pruning). The code and data are available athttps://github.com/lqsunshine/MCP.

Deep learning models are increasingly being employed in steganographic schemes for the embedding and extraction of secret information. However, steganographic models themselves are also at risk of detection and attacks. Although there are approaches proposed to hide deep learning models, making these models difficult to detect while achieving high-quality image steganography performance remains a challenging task. In this work, a robust image steganography method based on a camouflage model CamStegNet is proposed. The steganographic model is camouflaged as a routine deep learning model to significantly enhance its concealment. A sparse weight-filling paradigm is designed to enable the model to be flexibly switched among three modes by utilizing different keys: routine machine learning task, secret embedding task and secret recovery task. Furthermore, a residual state-space module and a neighborhood attention mechanism are constructed to improve the performance of image steganography. Experiments conducted on the DIV2K, ImageNet and COCO datasets demonstrate that the stego images generated by CamStegNet are superior to existing methods in terms of visual quality. They also exhibit enhanced resistance to steganalysis and maintain over 95% robustness against noise and scale attacks. Additionally, the model demonstrates high robustness which can achieve excellent performance in machine learning tasks and maintain stability across various weight initialization methods.

Cloud environments enhance diffusion model efficiency but introduce privacy risks, including intellectual property theft and data breaches. As AI-generated images gain recognition as copyright-protected works, ensuring their security and intellectual property protection in cloud environments has become a pressing challenge. This paper addresses privacy protection in diffusion model inference under cloud environments, identifying two key characteristics—denoising-encryption antagonism and stepwise generative nature—that create challenges such as incompatibility with traditional encryption, incomplete input parameter representation, and inseparability of the generative process. We propose PPIDM (Privacy-PreservingInference forDiffusionModels), a framework that balances efficiency and privacy by retaining lightweight text encoding and image decoding on the client while offloading computationally intensive U-Net layers to multiple non-colluding cloud servers. Client-side aggregation reduces computational overhead and enhances security. Experiments show PPIDM offloads 67% of Stable Diffusion computations to the cloud, reduces image leakage by 75%, and maintains high output quality (PSNR = 36.9, FID = 4.56), comparable to standard outputs. PPIDM offers a secure and efficient solution for cloud-based diffusion model inference.

Medical image description can be applied to clinical medical diagnosis, but the field still faces serious challenges. There is a serious problem of visual and textual data bias in medical datasets, which are the imbalanced distribution of health and disease data. This can greatly affect the learning performance of data-driven neural networks and finally lead to errors in the generated medical image descriptions. To address this problem, we propose a new medical image description network architecture named multimodal data-assisted knowledge fusion network (MDAKF), which introduces multimodal auxiliary signals to guide the Transformer network to generate more accurate medical reports. In detail, audio auxiliary signals provide clear abnormal visual regions to alleviate the visual data bias problem. However, the audio modality signals with similar pronunciation lack recognizability, which may lead to incorrect mapping of audio labels to medical image regions. Therefore, we further fuse the audio with text features as the auxiliary signal to improve the overall performance of the model. Through the experiments on two medical image description datasets, IU-X-ray and COV-CTR, it is found that the proposed model is superior to the previous models in terms of language generation evaluation indicators.

Existing coverless video steganography methods have not adequately exploited the stable features within and between video frames, and they have neglected the subtlety required for carrier transmission. To address these issues, this paper proposes a coverless video steganography method based on pose estimation and object tracking. By analyzing the intra-frame and inter-frame features of human posture within videos, this method hides secret information in videos depicting human activities, thereby enhancing concealment through simulating social behaviors. The scheme initially utilizes pose estimation network to localize target persons and their respective pose keypoints . Subsequently, a multi-object tracking algorithm is employed to track the detected targets within the video, coupled with a filtering mechanism to identify and prioritize tracking targets with larger areas, thus ensuring robustness in the tracking process. Then, corresponding hash mapping rules are established based on the inter-frame movement direction and the intra-frame angle features of the tracking targets . Finally, an inverted index is constructed to accelerate the speed of matching carrier videos containing the secret information and complete information hiding. Experimental results demonstrate that the proposed approach exhibits superior robustness against a variety of traditional attacks, video compression attacks, and frame dropping attacks compared to latest methods, while also enhancing the hiding capacity.

Frequent data breaches in the cloud environment have seriously affected cloud subscribers and providers. Privacy‐preserving image retrieval methods can improve the security of cloud image retrieval; however, existing methods have limited accuracy on dynamically updated image databases and mobile lightweight devices. In this study, we propose a privacy‐preserving image retrieval method based on disordered local histograms and vision transformer in cloud computing, by designing a multiple encryption method and transformer‐based feature model to better mine the local feature value of encrypted images. Specifically, the user performs different value substitution, position substitution, and color substitution on the subblocks of the image to protect the image information. The cloud server extracts the unordered local histogram from the encrypted image and generates retrievable features using transformer. Experiments show that compared with similar CNN schemes, the retrieval accuracy of this method is improved by 8.5%, and the retrieval efficiency is improved by 54.8%.

Recently, most coverless image steganography (CIS) methods are based on robust mapping rules. However, due to the limited mapping expression relationship between secret information and hash sequence, it is a challenge to further improve the hiding ability of coverless information hiding. Towards this goal, this paper proposes a robust coverless steganography scheme using camouflage image(CI-CIS). For the sender, CI-CIS introduces an camouflage image as the transmission carrier and establishes the correlation between them by Convolutional Neural Network(CNN) features. For the receiver, the camouflage image can retrieve the corresponding stego-image to recover the secret information. To this end, we designed a reversible retrieval scheme between stego-image and camouflage image by using image clustering. At the same time, since the semantic features represented by CNN are robust to image attacks, our method can increase the capability of the CIS effectively. Besides, we also build an inverted index to improve retrieval efficiency. Experimental results and analysis show that the CI-CIS has higher robustness and more flexible capacity setting compared with the existing CIS methods.

As more and more image data are stored in the encrypted form in the cloud computing environment, it has become an urgent problem that how to efficiently retrieve images on the encryption domain. Recently, Convolutional Neural Network (CNN) features have achieved promising performance in the field of image retrieval, but the high dimension of CNN features will cause low retrieval efficiency. Also, it is not suitable to directly apply them for image retrieval on the encryption domain. To solve the above issues, this paper proposes an improved CNN-based hashing method for encrypted image retrieval. First, the image size is increased and inputted into the CNN to improve the representation ability. Then, a lightweight module is introduced to replace a part of modules in the CNN to reduce the parameters and computational cost. Finally, a hash layer is added to generate a compact binary hash code. In the retrieval process, the hash code is used for encrypted image retrieval, which greatly improves the retrieval efficiency. The experimental results show that the scheme allows an effective and efficient retrieval of encrypted images.

With the rapid development of interactive multimedia services and camera sensor networks, the number of network videos is exploding, which has formed a natural carrier library for steganography. In this study, a coverless steganography scheme based on motion analysis of video is proposed. For every video in the database, the robust histograms of oriented optical flow (RHOOF) are obtained, and the index database is constructed. The hidden information bits are mapped to the hash sequences of RHOOF, and the corresponding indexes are sent by the sender. At the receiver, through calculating hash sequences of RHOOF from the cover video, the secret information can be extracted successfully. During the whole process, the cover video remains original without any modification and has a strong ability to resist steganalysis. The capacity is investigated and shows good improvement. The robustness performance is prominent against most attacks such as pepper and salt noise, speckle noise, MPEG-4 compression, and motion JPEG 2000 compression. Compared with the existing coverless information hiding schemes based on images, the proposed method not only obtains a good trade-off between hiding information capacity and robustness but also can achieve higher hiding success rate and lower transmission data load, which shows good practicability and feasibility.

Most of the existing coverless steganography approaches have poor robustness to geometric attacks, because these approaches use features of the entire image to map information, and these features are easy to be lost when being attacked. In order to improve the robustness against geometric attacks, we propose a coverless image steganography method based on multi-object recognition. In this scheme, we firstly use Faster RCNN to detect objects in the image data set, establish a mapping dictionary between object labels and binary sequence. Then we propose a novel mapping rule based on the filtered robust object labels for sequence generation. Therefore, an image can generate robust binary sequence through multi-objects recognition. In the transmission process, the transmitted image has not been modified, so our method can fundamentally resist steganalysis tools and avoid the attacker’s suspicions. In addition, the capacity and hiding rate of the proposed method are both considerable. Evaluations with under geometric attacks shows, on average,$3.1\times $robustness increase over other five coverless steganography methods. Moreover, evaluations under ten noise attacks shows, on average, the robustness of our method is also excellent, which reaches 83%.