Search Results (21)

Large-scale Visual-Language Models have demonstrated powerful adaptability in video recognition tasks. However, existing methods typically rely on fine-tuning or text prompt tuning. In this paper, we propose a visual-only prompting method that employs watermark and trademark prompts to bridge the distribution gap of spatial-temporal video data with Visual-Language Models. Our watermark prompts, designed by a trainable prompt generator, are customized for each video clip. Unlike conventional visual prompts that often exhibit noise signals, watermark prompts are intentionally designed to be imperceptible, ensuring they are not misinterpreted as an adversarial attack. The trademark prompts, bespoke for each video domain, establish the identity of specific video types. Integrating watermark prompts into video frames and prepending trademark prompts to per-frame embeddings significantly boosts the capability of the Visual-Language Model to understand video. Notably, our approach improves the adaptability of the CLIP model to various video action recognition datasets, achieving performance gains of 16.8%, 18.4%, and 13.8% on HMDB-51, UCF-101, and the egocentric dataset EPIC-Kitchen-100, respectively. Additionally, our visual-only prompting method demonstrates competitive performance compared with existing fine-tuning and adaptation methods while requiring fewer learnable parameters. Moreover, through extensive ablation studies, we find the optimal balance between imperceptibility and adaptability. Code will be made available. Full article

With the exponential growth in transactions and exchanges carried out via the Internet, the risks of the falsification and distortion of information are multiplying, encouraged by widespread access to the virtual world. In this context, digital image watermarking has emerged as an essential solution for protecting digital content by enhancing its durability and resistance to manipulation. However, no current digital watermarking technology offers complete protection against all forms of attack, with each method often limited to specific applications. This field has recently benefited from the integration of deep learning techniques, which have brought significant advances in information security. This article explores the implementation of digital watermarking in embedded systems, addressing the challenges posed by resource constraints such as memory, computing power, and energy consumption. We propose optimization techniques, including frequency domain methods and the use of lightweight deep learning models, to enhance the robustness and resilience of embedded systems. The experimental results validate the effectiveness of these approaches for enhanced image protection, opening new prospects for the development of information security technologies adapted to embedded environments. Full article

As generative AI (GenAI) technologies proliferate, ensuring trust and transparency in digital ecosystems becomes increasingly critical, particularly within democratic frameworks. This article examines decentralized Web3 mechanisms—blockchain, decentralized autonomous organizations (DAOs), and data cooperatives—as foundational tools for enhancing trust in GenAI. These mechanisms are analyzed within the framework of the EU’s AI Act and the Draghi Report, focusing on their potential to support content authenticity, community-driven verification, and data sovereignty. Based on a systematic policy analysis, this article proposes a multi-layered framework to mitigate the risks of AI-generated misinformation. Specifically, as a result of this analysis, it identifies and evaluates seven detection techniques of trust stemming from the action research conducted in the Horizon Europe Lighthouse project called ENFIELD: (i) federated learning for decentralized AI detection, (ii) blockchain-based provenance tracking, (iii) zero-knowledge proofs for content authentication, (iv) DAOs for crowdsourced verification, (v) AI-powered digital watermarking, (vi) explainable AI (XAI) for content detection, and (vii) privacy-preserving machine learning (PPML). By leveraging these approaches, the framework strengthens AI governance through peer-to-peer (P2P) structures while addressing the socio-political challenges of AI-driven misinformation. Ultimately, this research contributes to the development of resilient democratic systems in an era of increasing technopolitical polarization. Full article

The tasks assigned to neural network (NN) models are increasingly challenging due to the growing demand for their applicability across domains. Advanced machine learning programming skills, development time, and expensive assets are required to achieve accurate models, and they represent important assets, particularly for small and medium enterprises. Whether they are deployed in the Cloud or on Edge devices, i.e., resource-constrained devices that require the design of tiny NNs, it is of paramount importance to protect the associated intellectual properties (IP). Neural networks watermarking (NNW) can help the owner to claim the origin of an NN model that is suspected to have been attacked or copied, thus illegally infringing the IP. Adapting two state-of-the-art NNW methods, this paper aims to define watermarking procedures to securely protect tiny NNs’ IP in order to prevent unauthorized copies of these networks; specifically, embedded applications running on low-power devices, such as the image classification use cases developed for MLCommons benchmarks. These methodologies inject into a model a unique and secret parameter pattern or force an incoherent behavior when trigger inputs are used, helping the owner to prove the origin of the tested NN model. The obtained results demonstrate the effectiveness of these techniques using AI frameworks both on computers and MCUs, showing that the watermark was successfully recognized in both cases, even if adversarial attacks were simulated, and, in the second case, if accuracy values, required resources, and inference times remained unchanged. Full article

This work and manuscript focus on how 3D scanning methodologies and post-processing analyses may help us to gain a deeper investigation of cuneiform tablets beyond the written content. The dataset proposed herein is a key part of the archaeological collection preserved in the Musei Reali of Turin in Italy; these archaeological artefacts enclose further important semantic information extractable through detailed 3D documentation and 3D model filtering. In fact, this scanning process is a fundamental tool for better reading of sealing impressions beneath the cuneiform text, as well as for understanding micrometric evidence of the fingerprints of scribes. Most of the seal impressions were made before the writing (like a watermark), and thus, they are not detectable to the naked eye due to cuneiform signs above them as well as the state of preservation. In this regard, 3D scanning and post-processing analysis could help in the analysis of these nearly invisible features impressed on tablets. For this reason, this work is also based on how 3D analyses may support the identification of the unperceived and almost invisible features concealed in clay tablets. Analysis of fingerprints and the depths of the signs can tell us about the worker’s strategies and the people beyond the artefacts. Three-dimensional models generated inside the Artec 3D ecosystem via Space Spider scanner and Artec Studio software were further investigated by applying specific filters and shaders. Digital light manipulation can reveal, through the dynamic displacement of light and shadows, particular details that can be deeply analysed with specific post-processing operations: for example, the MSII (multi-scale integral invariant) filter is a powerful tool exploited for revealing hidden and unperceived features such as fingerprints and sealing impressions (stratigraphically below cuneiform signs). Finally, the collected data will be handled twofold: in an open-access repository and through a common data environment (CDE) to aid in the data exchange process for project collaborators and common users. Full article

To ensure the security of online-shared power data, this paper adopts a robust soliton distribution-based zero-watermarking approach for tracing semi-structured power data. The method involves extracting partial key-value pairs to generate a feature sequence, processing the watermark into an equivalent number of blocks. Robust soliton distribution from erasure codes and redundant error correction codes is utilized to generate an intermediate sequence. Subsequently, the error-corrected watermark information is embedded into the feature sequence, creating a zero-watermark for semi-structured power data. In the tracking process, the extraction and analysis of the robust zero-watermark associated with the tracked data facilitate the effective identification and localization of data anomalies. Experimental and simulation validation demonstrates that this method, while ensuring data security, achieves a zero-watermark extraction success rate exceeding 98%. The proposed approach holds significant application value for data monitoring and anomaly tracking in power systems. Full article

In order to safeguard image copyrights, zero-watermarking technology extracts robust features and generates watermarks without altering the original image. Traditional zero-watermarking methods rely on handcrafted feature descriptors to enhance their performance. With the advancement of deep learning, this paper introduces “ZWNet”, an end-to-end zero-watermarking scheme that obviates the necessity for specialized knowledge in image features and is exclusively composed of artificial neural networks. The architecture of ZWNet synergistically incorporates ConvNeXt and LK-PAN to augment the extraction of local features while accounting for the global context. A key aspect of ZWNet is its watermark block, as the network head part, which fulfills functions such as feature optimization, identifier output, encryption, and copyright fusion. The training strategy addresses the challenge of simultaneously enhancing robustness and discriminability by producing the same identifier for attacked images and distinct identifiers for different images. Experimental validation of ZWNet’s performance has been conducted, demonstrating its robustness with the normalized coefficient of the zero-watermark consistently exceeding 0.97 against rotation, noise, crop, and blur attacks. Regarding discriminability, the Hamming distance of the generated watermarks exceeds 88 for images with the same copyright but different content. Furthermore, the efficiency of watermark generation is affirmed, with an average processing time of 96 ms. These experimental results substantiate the superiority of the proposed scheme over existing zero-watermarking methods. Full article

Natural language generation (NLG) models combined with increasingly mature and powerful deep learning techniques have been widely used in recent years. Deployed NLG models in practical applications may be stolen or used illegally, and watermarking has become an important tool to protect the Intellectual Property (IP) of these deep models. Watermarking technique designs algorithms to embed watermark information and extracts watermark information for IP identification of NLG models can be seen as a symmetric signal processing problem. In terms of IP protection of NLG models, however, the existing watermarking approaches cannot provide reliable and timely model protection and prevent illegal users from utilizing the original performance of the stolen models. In addition, the quality of watermarked text sequences generated by some watermarking approaches is not high. In view of these, this paper proposes two embedding schemes to the hidden memory state of the RNN to protect the IP of NLG models for different tasks. Besides, we add a language model loss to the model decoder to improve the grammatical correctness of the output text sequences. During the experiments, it is proved that our approach does not compromise the performance of the original NLG models on the corresponding datasets and outputs high-quality text sequences, while forged secret keys will generate unusable NLG models, thus defeating the purpose of model infringement. Besides, we also conduct sufficient experiments to prove that the proposed model has strong robustness under different attacks. Full article

Image steganography is a subfield of pattern recognition. It involves hiding secret data in a cover image and extracting the secret data from the stego image (described as a container image) when needed. Existing image steganography methods based on Deep Neural Networks (DNN) usually have a strong embedding capacity, but the appearance of container images is easily altered by visual watermarks of secret data. One of the reasons for this is that, during the end-to-end training process of their Hiding Network, the location information of the visual watermarks has changed. In this paper, we proposed a layerwise adversarial training method to solve the constraint. Specifically, unlike other methods, we added a single-layer subnetwork and a discriminator behind each layer to capture their representational power. The representational power serves two purposes: first, it can update the weights of each layer which alleviates memory requirements; second, it can update the weights of the same discriminator which guarantees that the location information of the visual watermarks remains unchanged. Experiments on two datasets show that the proposed method significantly outperforms the most advanced methods. Full article

As a self-supervised learning paradigm, contrastive learning has been widely used to pre-train a powerful encoder as an effective feature extractor for various downstream tasks. This process requires numerous unlabeled training data and computational resources, which makes the pre-trained encoder become the valuable intellectual property of the owner. However, the lack of a priori knowledge of downstream tasks makes it non-trivial to protect the intellectual property of the pre-trained encoder by applying conventional watermarking methods. To deal with this problem, in this paper, we introduce AWEncoder, an adversarial method for watermarking the pre-trained encoder in contrastive learning. First, as an adversarial perturbation, the watermark is generated by enforcing the training samples to be marked to deviate respective location and surround a randomly selected key image in the embedding space. Then, the watermark is embedded into the pre-trained encoder by further optimizing a joint loss function. As a result, the watermarked encoder not only performs very well for downstream tasks, but also enables us to verify its ownership by analyzing the discrepancy of output provided using the encoder as the backbone under both white-box and black-box conditions. Extensive experiments demonstrate that the proposed work enjoys quite good effectiveness and robustness on different contrastive learning algorithms and downstream tasks, which has verified the superiority and applicability of the proposed work. Full article

The reconfigurable feature of FPGAs (Field-Programmable Gate Arrays) has made them a very attractive solution for implementing adaptive systems-on-chip. However, this implies additional design tasks to handle system reconfiguration and control, which increases design complexity. To address this issue, this paper proposes a model-driven design flow that guides the designer through the description of the different elements of a reconfigurable system. It is based on high-level modeling using an extended version of the MARTE (Modeling and Analysis of Real-Time and Embedded systems) UML (Unified Modeling Language) profile. Both centralized and decentralized reconfiguration decision-making solutions are possible with the proposed flow, allowing it to adapt to various reconfigurable systems constraints. It also integrates the IP-XACT standard (standard for the description of electronic Intellectual Properties), allowing the designer to easily target different technologies and commercial FPGAs by reusing both high-level models and actual IP-XACT hardware components. At the end of the flow, the implementation code is generated automatically from the high-level models. The proposed design flow was validated through a reconfigurable video watermarking application as a case study. Experimental results showed that the generated system allowed a good trade-off between resource usage, power consumption, execution time, and image quality compared to static implementations. This hardware efficiency was achieved in a very short time thanks to the design acceleration and automation offered by model-driven engineering. Full article

Various hardware security concerns, such as hardware Trojans and IP piracy, have sparked studies in the security field employing alternatives to CMOS chips. Spintronic devices are among the most-promising alternatives to CMOS devices for applications that need low power consumption, non-volatility, and ease of integration with silicon substrates. This article looked at how hardware can be made more secure by utilizing the special features of spintronics devices. Spintronic-based devices can be used to build polymorphic gates (PGs), which conceal the functionality of the circuits during fabrication. Since spintronic devices such as magnetic tunnel junctions (MTJs) offer non-volatile properties, the state of these devices can be written only once after fabrication for correct functionality. Symmetric circuits using two-terminal MTJs and three-terminal MTJs were designed, analyzed, and compared in this article. The simulation results demonstrated how a single control signal can alter the functionality of the circuit, and the adversary would find it challenging to reverse-engineer the design due to the similarity of the logic blocks’ internal structures. The use of spintronic PGs in IC watermarking and fingerprinting was also explored in this article. The TSMC 65nm MOS technology was used in the Cadence Spectre simulator for all simulations in this work. For the comparison between the structures based on different MTJs, the physical dimension of the MTJs were kept precisely the same. Full article

Power noise generation for masking power traces is a powerful countermeasure against Simple Power Analysis (SPA), and it has also been used against Differential Power Analysis (DPA) or Correlation Power Analysis (CPA) in the case of cryptographic circuits. This technique makes use of power consumption generators as basic modules, which are usually based on ring oscillators when implemented on FPGAs. These modules can be used to generate power noise and to also extract digital signatures through the power side channel for Intellectual Property (IP) protection purposes. In this paper, a new power consumption generator, named Xored High Consuming Module (XHCM), is proposed. XHCM improves, when compared to others proposals in the literature, the amount of current consumption per LUT when implemented on FPGAs. Experimental results show that these modules can achieve current increments in the range from 2.4 mA (with only 16 LUTs on Artix-7 devices with a power consumption density of 0.75 mW/LUT when using a single HCM) to 11.1 mA (with 67 LUTs when using 8 XHCMs, with a power consumption density of 0.83 mW/LUT). Moreover, a version controlled by Pulse-Width Modulation (PWM) has been developed, named PWM-XHCM, which is, as XHCM, suitable for power watermarking. In order to build countermeasures against SPA attacks, a multi-level XHCM (ML-XHCM) is also presented, which is capable of generating different power consumption levels with minimal area overhead (27 six-input LUTS for generating 16 different amplitude levels on Artix-7 devices). Finally, a randomized version, named RML-XHCM, has also been developed using two True Random Number Generators (TRNGs) to generate current consumption peaks with random amplitudes at random times. RML-XHCM requires less than 150 LUTs on Artix-7 devices. Taking into account these characteristics, two main contributions have been carried out in this article: first, XHCM and PWM-XHCM provide an efficient power consumption generator for extracting digital signatures through the power side channel, and on the other hand, ML-XHCM and RML-XHCM are powerful tools for the protection of processing units against SPA attacks in IoT devices implemented on FPGAs. Full article

In this paper, a color video watermarking system based on SVD in the complex wavelet domain is proposed. The process of inserting copyright information in video bit streams is known as video watermarking. It has been advocated in recent years as a solution to the problem of unlawful digital video alteration and dissemination. An effective, robust, and invisible video watermarking algorithm is proposed in this paper. The two-level dual tree complex wavelet transform (DT-CWT) and singular value decomposition are used to create this approach, which was built on a cascade of two powerful mathematical transforms. This hybrid technique demonstrates a high level of security as well as various levels of attack robustness. The proposed algorithm was used to the test for imperceptibility and robustness, and this resulted in excellent grades. We compared our suggested method to a DWT-SVD-based technique and found it to be far more reliable and effective. Full article

In the past years, Deep Neural Networks (DNNs) have become popular in many disciplines such as Computer Vision (CV), and the evolution of hardware has helped researchers to develop many powerful Deep Learning (DL) models to deal with several problems. One of the most important challenges in the CV area is Medical Image Analysis. However, adversarial attacks have proven to be an important threat to vision systems by significantly reducing the performance of the models. This paper brings to light a different side of digital watermarking, as a potential black-box adversarial attack. In this context, apart from proposing a new category of adversarial attacks named watermarking attacks, we highlighted a significant problem, as the massive use of watermarks, for security reasons, seems to pose significant risks to vision systems. For this purpose, a moment-based local image watermarking method is implemented on three modalities, Magnetic Resonance Images (MRI), Computed Tomography (CT-scans), and X-ray images. The introduced methodology was tested on three state-of-the art CV models, DenseNet 201, DenseNet169, and MobileNetV2. The results revealed that the proposed attack achieved over 50% degradation of the model’s performance in terms of accuracy. Additionally, MobileNetV2 was the most vulnerable model and the modality with the biggest reduction was CT-scans. Full article