Electronic evidence of forensic medical images plays a key role in forensic identification. The existing deposit technology is difficult to cope with the dual challenges of image format change and AI forgery, and the fusion mechanism of digital watermarking and blockchain has the problems of robustness and traceability accuracy imbalance. This article proposes a dynamic trusted certificate storage system that integrates deep learning perceptual hash and alliance chain. A semantic hash generation network based on multi-scale frequency domain features is designed, and a lightweight intelligent contract architecture optimized by SM2/SM3 algorithm of state secrets is established. The full link traceability is realized by combining adaptive frequency domain and time domain nested watermarking algorithms. Experiments show that under the attacks of Gaussian noise, JPEG compression and geometric deformation, the Hamming distance of the hash is stable within 3 bits, which is better than the mutation of more than 30 bits in the traditional cryptographic hash. When the rotation is 10, the false recognition rate is less than 1%, and the sample collision probability is maintained at a very low order of magnitude; When the watermark embedding strength increases, the PSNR remains above 45 dB, and the normalized cross-correlation coefficient is higher than 0.92 under the condition of JPEG compression quality of 70. In the alliance chain scenario, the consensus delay of 30 nodes is 180 ms, and the delay rises to 320ms after the expansion of 100 nodes, and the system throughput is not significantly attenuated. In this study, the synergy between robustness, transparency and traceability efficiency is optimized, which can provide a reference technical scheme for judicial acceptance of forensic electronic evidence chain.
Suman Bijapur, Shilpa Patil, Parimala, Shantala P H
With unparalleled threats to the integrity of digital information, democratic practices, and public confidence in media, deepfake technology comprises a new class of harm. Deep generative models can be used to generate realistic looking (and sounding) fake human faces and voices, which is great news for bad actors who seek to spread misinformation, commit crimes and ruin journalism. CyberLink Fights Deepfakes with New AI Model That Uses Neural Network Traditional methods used to identify deepfakes have depended on centralised AI systems that can't be trusted at face value and there is little or no way of proving a piece of content's authenticity. In this work, we have presented a solution that involves multi-modal deepfake detection and has utilized learning-based forgery detection framework to be deployed on blockchain for evidence tamper resistance. The proposed framework employs a hybrid CNN-RNN architecture that computes facial, audio and metadata feature in parallel to detect unseen deepfakes with accuracy of 94.2%, compared to the single-modal baselines (CNN only: 81.3%, and audio only: 67.4%). Novelty: Blockchain timestamping with cryptographically secured certificates of authenticity for third-party verification while protected proprietary detection logic is not revealed. At the computational efficiency and bandwidth threshold required for edge deployment, video processing at 30 FPS and only 2.1 MBs makes this applicable on any average mobile device. It holds for 12k synthetic videos (celebrities, politicians, newscasters) and diverse deepfake generation methods (FaceSwap, DeepFaceLab, StyleGAN). Societal impact: framework mitigates $1.2T annual disinformation damage and champions digital rights through decentralized verification. Via mashable.com Framework addresses the convergence of deepfake detection, blockchain authentication and the case for sustainable cybersecurity: As a global community grapples with synthetic media in ways we've never seen before, support online safety experts to respond.
Open access
2 source records
Generative Adversarial Networks and Image Synthesis
Ensuring the authenticity, integrity, and reliability of digital image evidence is a persistent challenge in forensic and legal domains due to the vulnerabilities of centralized evidence management systems. This study compares three blockchain consensus mechanisms—Proof of Existence (PoE), Proof of Ownership (PoOW), and Zero-Knowledge Ethereum Virtual Machine (zkEVM)—to assess their effectiveness in securing and validating digital image evidence on the Layer 2 Polygon network. Forensic images were stored on the InterPlanetary File System (IPFS), with each consensus model registering tamper-evident Content Identifiers (CIDs) on-chain via dedicated smart contracts. The evaluation considered performance metrics including latency, gas usage, transaction fees, throughput, scalability, and privacy protection. The findings revealed that PoE demonstrated the best overall efficiency, achieving a latency of 3730ms, a transaction fee of 0.001321 ETH, and a throughput of 0.176 TPS, making it well-suited for real-time applications such as timestamping and immediate evidence submission. PoOW, although more computationally demanding, achieved the highest gas-refund rate at 89%, making it ideal for ownership verification and traceability, such as copyright and asset provenance. Meanwhile, zkEVM provided a well-rounded performance profile with moderate transaction costs and latency. It is powerful for privacy-preserving applications that require cryptographic guarantees, especially in enterprise and regulatory settings. This comparative evaluation highlights the unique advantages and limitations of each approach, providing critical insights into selecting the most suitable blockchain-based consensus mechanism for the transparent and tamper-resistant validation of digital forensic evidence.
Democratic elections rely on trust, transparency, and tamper-resistance -- qualities that conventional and early electronic voting systems have consistently failed to guarantee. This paper presents a Blockchain-Enabled Secure E-Voting Framework with Facial Recognition for Voter Authentication, designed to address persistent vulnerabilities in existing electoral systems. The proposed system integrates a permissioned blockchain ledger with deep-learning-based facial biometric verification to ensure decentralized, immutable vote storage and strong identity assurance. A multi-layer security architecture combines homomorphic encryption, zero-knowledge proofs, and digital signatures to preserve voter anonymity while enabling end-to-end verifiability. Anti-spoofing and liveness detection mechanisms prevent impersonation via photographs, video replays, or deepfake-generated imagery. Smart contracts automate vote counting and result publication, eliminating human involvement in the tallying process. Experimental evaluation demonstrates a facial recognition authentication accuracy of 97.3% and an end-to-end voting transaction latency under 500 milliseconds, with blockchain confirmation averaging 2.4 seconds. The framework is scalable to national-scale elections and applicable to governmental, corporate, and institutional governance contexts.
Abstract AI-generated forgeries of financial documents—such as invoices, audit reports, ledgers, and balance sheets—expose a critical fault line in legal proof. These hybrid visual–textual artefacts derive evidentiary authority from their jurisvisual form: logos, seals, signatures, and tabular architecture, whose visual grammar indexes authenticity and institutional power. Drawing on Charles Sanders Peirce’s triadic semiotics (representamen–object–interpretant), this study demonstrates that deepfake technologies dissolve the sign-relation underwriting documentary proof by engineering synthetic representamina that mimic the indexical and symbolic features of authentic documents. At the same time, the underlying financial event may be absent. The evidentiary economy is thereby reconfigured within a videosphere where image-like documents perform the truth. This article advances a layered remediation architecture: (i) provenance anchoring through cryptographic signatures, content hashing, and distributed ledgers; (ii) content forensics integrating AI-assisted detection with forensic semiotics—indexical stress tests and symbolic authenticity challenges; and (iii) procedural safeguards including calibrated evidentiary thresholds, adversarial authenticity hearings, and robust chain-of-custody protocols. It argues that restoring evidentiary confidence requires cultivating semiotic literacy among judges, auditors, and legal practitioners as core professional competence, enabling legal systems to navigate the post-textual landscape with epistemic rigour.
Gustav Olaf Yunus Laitinen-Fredriksson Lundström-Imanov, N I Abdullayeva
Generative artificial intelligence now synthesizes photorealistic imagery, audio, and video at a cost that defeats traditional forensic intuition. The legal consequences span three regimes studied so far in isolation: international operational law, domestic procedure, and product regulation. This article presents a unified evidentiary framework that maps cryptographic content provenance, robust statistical watermarking, and zero knowledge attestation to the proof requirements of each regime. We define a five tier threat model spanning naive regeneration, adversarial laundering, cross model regeneration, active watermark removal, and insider provenance forgery. We release a public benchmark of 12000 generated items across image, audio, and video modalities under six laundering pipelines for 72000 evaluation samples. We evaluate four representative schemes and report true positive rate at fixed false positive rate, robustness area under the curve, computational overhead, and a regime conditioned legal sufficiency score. We translate empirical detection bounds into legal sufficiency thresholds for command decisions under the law of armed conflict, for criminal and civil admissibility under domestic procedure, and for persistence audits under the European Union Artificial Intelligence Act and analogous regimes. The result is a reproducible reference pipeline, a public benchmark, and model annexes that lawyers, engineers, and operators can deploy together.
Open access
2 source records
Adversarial Robustness in Machine Learning
Digital Media Forensic Detection
Generative Adversarial Networks and Image Synthesis
As neural language models are deployed in regulated domains, verifiable model provenance becomes a critical security requirement. We construct an Inference-Time Physical Unclonable Function (IT-PUF) that provides a challenge-response authentication protocol for neural networks, achieving zero false acceptances across 1,012 comparisons spanning 23 models and 16 vendor families. The IT-PUF derives its entropy from a geometrically intrinsic behavioral fingerprint—the delta-gene (the third pre-softmax logit gap)—which we prove is invariant to inference temperature and empirically validate as invariant across six distinct neural architectures. We provide a formal impossibility result for fingerprint spoofing: an interval-splitting theorem proves that no adversarial Kullback-Leibler (KL) budget can simultaneously close the fingerprint gap and avoid detection via accumulated noise. To establish that this security does not degrade at scale, we validate an Equation of State across three independent model families spanning a 147x parameter range (0.5B to 72B). We falsify the assumption of unbounded stiffness but discover a strict positive empirical floor (S_min = 1.1797), from which the Cramér-Rao bound guarantees a computable minimum spoofing cost. The theoretical foundation is formally verified in the Coq proof assistant: 311 theorems across 16 files, with zero uses of "Admitted" and zero vacuous definitions. The Neural Network Identity Series — Mathematical foundations, empirical validation, and governance frameworks for verifying which model is running Newest addition: Technical Note: The Disappearing Window — AI Logprob Access Withdrawal and the Structural Verifiability of Frontier Model Contracts (DOI: 10.5281/zenodo.20362098) Paper 1: The δ-Gene: Inference-Time Physical Unclonable Functions from Architecture-Invariant Output Geometry (DOI: 10.5281/zenodo.18704275) Paper 2: Template-Based Endpoint Verification via Logprob Order-Statistic Geometry (DOI: 10.5281/zenodo.18776711) Paper 3: The Geometry of Model Theft: Distillation Forensics, Adversarial Erasure, and the Illusion of Spoofing (DOI: 10.5281/zenodo.18818608) Paper 4: Provenance Generalization and Verification Scaling for Neural Network Forensics (DOI: 10.5281/zenodo.18872071) Paper 5: Beneath the Character: The Structural Identity of Neural Networks — Mathematical Evidence for a Non-Narrative Layer of AI Identity (DOI: 10.5281/zenodo.18907292) Paper 6: Which Model Is Running?: Structural Identity as a Prerequisite for Trustworthy Zero-Knowledge Machine Learning (DOI: 10.5281/zenodo.19008116) Paper 7: The Deformation Laws of Neural Identity (DOI: 10.5281/zenodo.19055966) Paper 8: What Counts as Proof? — Admissible Evidence for Neural Network Identity Claims (DOI: 10.5281/zenodo.19058540) Paper 9: Composable Model Identity — Formal Hardening of Structural Attestations in the Enterprise Identity Stack (DOI: 10.5281/zenodo.19099911) Paper 10:Where Identity Comes From: Path Sensitivity and Endpoint Underdetermination in Neural Network Training (DOI: 10.5281/zenodo.19118807) Paper 11: Post-Hoc Disclosure Is Not Runtime Proof: Model Identity at Frontier Scale (DOI: 10.5281/zenodo.19216634) Paper 12: Family-Dependent Response to Reasoning Distillation Across Structural and Functional Identity Layers (DOI: 10.5281/zenodo.19298857) Paper 13: Safety-Alignment Removal as a Model-Identity Failure — Structural Evidence from Published Weight-Level Mutation Checkpoints (DOI: 10.5281/zenodo.19383019) Technical Note: Agent Identity Is Not Model Identity (DOI: 10.5281/zenodo.19240883) Technical Note: Gap Invariance: Why PPP Measurements Are Domain-Independent by Construction (DOI: 10.5281/zenodo.19275524) Technical Note: Measured Model Substitution Under Valid Agent Credentials (DOI: 10.5281/zenodo.19342848) Technical Note: Artifact Identity Is Not Runtime Identity — Trustfall Lite and the Boundary of File-Level Model Verification (DOI: 10.5281/zenodo.20019127) Formal Verification Stack for Neural Network Structural Identity (IT-PUF Coq Proofs) (DOI: 10.5281/zenodo.18930621) Copyright (c) 2026 Anthony Ray Coslett / Fall Risk AI, LLC. All Rights Reserved. Confidential and Proprietary. Patent Pending (Applications 63/982,893, 63/990,487, 63/996,680, 64/003,244).
Open access
2 source records
Physical Unclonable Functions (PUFs) and Hardware Security
Adversarial Robustness in Machine Learning
Generative Adversarial Networks and Image Synthesis
Within the context of digital forensics, the integrity and authenticity of digital evidence are crucial for its legal admissibility within a courtroom setting. Chain of Custody (CoC) processes ensure that digital evidence is meticulously managed and documented from its point of origin until its use in legal proceedings. As the importance of digital forensics increases, especially with cybercrime investigations, the traditional processes used in traditional Chain of Custody have challenges in terms of transparency, security, and efficiency. This paper highlights some of the recent developments in Chain of Custody processes, particularly with the adoption of blockchain and Artificial Intelligence technologies. Blockchain technology, known for its impenetrable and distributed properties, introduces a new paradigm for Chain of Custody processes, enhancing security and traceability for digital evidence management. Additionally, AI-based algorithms for anomaly detection have the potential for increasing the reliability of Chain of Custody processes. Moreover, we will explore the decentralized evidence storage approaches and privacy-preserving mechanisms, such as zero-knowledge proofs. These are important in ensuring that more secure yet transparent approaches in managing distributed forensic investigation systems are achieved. The effectiveness of currently used CoC approaches presents lessons in understanding the future of improving the integrity of this process. Such innovations have the potential of revolutionizing the field of digital forensic investigation processes while ensuring that the handling of such evidence is of the highest integrity.
In the digital era, protecting visual content from misuse and forgery is essential. This study proposes a robust image watermarking method by integrating Discrete Wavelet Transform (DWT), Hessenberg Decomposition (HD), and Singular Value Decomposition (SVD), aiming to enhance watermark imperceptibility and resilience against common image attacks. Additionally, the system incorporates RSA digital signatures within the watermark metadata to ensure verifiable authenticity in NFT (Non-Fungible Token) applications. The method was implemented using Python and tested on multiple grayscale images across various attack scenarios, including noise addition and compression. Experimental results demonstrate high SSIM and PSNR values, confirming the method's effectiveness in maintaining both visual fidelity and embedded watermark integrity. These findings support the potential of this approach for secure and scalable NFT copyright protection.
Open access
Advanced Steganography and Watermarking Techniques
The credibility of digital evidence is a cornerstone of modern cybercrime investigations, digital forensics, and judicial processes. However, adversarial tampering, deepfake manipulation, and insider threats have raised significant concerns regarding the authenticity and admissibility of such evidence. Conventional integrity-preservation methods—such as hashing, encryption, and secure storage—struggle to meet the demands of scalability, transparency, and resilience in today’s forensic environments. Recent advances in artificial intelligence (AI) and blockchain offer promising avenues for overcoming these limitations. AI techniques contribute to content-level verification by detecting anomalies, forgeries, and manipulations in digital artefacts, while blockchain ensures tamper-proof chain-of-custody management through decentralization, immutability, and auditability. This review synthesizes the state of the art in digital evidence integrity verification through the combined application of AI and blockchain. We examine existing frameworks, datasets, algorithms, and deployment models, while critically analyzing their strengths and limitations. Furthermore, we identify gaps in scalability, explainability, and legal admissibility, proposing future directions such as federated learning, explainable AI, zero-knowledge proofs, and quantum-resistant blockchains. By consolidating research across computer science, law, and digital forensics, this review highlights the potential of AI–blockchain synergy to establish robust, scalable, and trustworthy evidence verification frameworks for real-world forensic and judicial systems.
We propose a novel blockchain-based traceability system that uniquely combines Physical Unclonable Functions (PUFs) and Non-Fungible Tokens (NFTs) to establish secure, tamper-evident, and verifiable digital identities for physical products. Unlike conventional approaches that rely solely on serial numbers or barcodes, our system uses embedded PUFs to generate a physically unclonable ID for each item, ensuring hardware-level authenticity. This PUF ID is cryptographically linked to an NFT minted on a public blockchain, encapsulating metadata such as product origin, manufacturing details, and certification status. Associating NFTs and certifications with PUF-tagged products not only provides transparent provenance but also enables decentralised validation of compliance and ethical standards. This methodology is distinct in its integration of immutable physical identity with blockchain-based digital certification, offering a robust solution for enhancing transparency, combating counterfeiting, and fostering trust across global supply chains. Furthermore, employing Edge and emerging 6G networks, our framework can execute PUF challenge–response and preliminary NFT minting directly at edge nodes. This establishes a real-time, low-latency architecture that reduces network congestion and gas costs for secure and sustainable supply chains.
Open access
Physical Unclonable Functions (PUFs) and Hardware Security
With the rapid proliferation of smart home cameras, wearable vision devices, and user-generated Consumer Internet of Things (CIoT) content, ensuring visual media authenticity, rightful ownership, and tamper detection has become increasingly challenging. We propose Chain-Visage, a blockchain-assisted framework for secure content authentication and tamper tracing in decentralized CIoT multimedia ecosystems. The termChainreflects the consortium blockchain backbone that guarantees immutable provenance, decentralized ownership management, and copyright revocation, whileVisagesymbolizes the unique visual identity of multimedia content achieved through dual-stage visual hash embedding. The proposed framework integrates Zero-Knowledge Proofs (ZKPs) for privacy-preserving ownership verification and employs optimized smart contracts to manage visual rights, provenance records, and ownership transfers efficiently. Evaluations on a large-scale dataset of over 10,000 real-world images and 3,850 video clips from diverse CIoT devices demonstrate Chain-Visage’s superior performance, achieving 97.5% traceability accuracy, 93% tamper detection sensitivity, and low verification latency even under resource-constrained environments. This work addresses a critical research gap in secure, privacy-preserving, and energy-efficient multimedia ownership control and tamper-resilient content authentication for next-generation CIoT ecosystems.
Open access
Blockchain Technology Applications and Security
Digital Media Forensic Detection
Advanced Steganography and Watermarking Techniques
Digital forensic investigation in 2025 faces unprecedented challenges posed by the convergence of decentralized web technologies (Web3), adversarial generative AI systems, and darknet infrastructure. Traditional attribution and evidence preservation methodologies prove in-sufficient when adversaries exploit blockchain immutability, synthetic media generation, and privacy-enhancing technologies to obscure malicious intent. This paper in-traduces SHARD (Shadowed and Silicon Hybrid Attribution and Reconstruction Diagnostic), a multi-modal forensic framework designed to recover, correlate, and at-tribute malicious artifacts across distributed ledger systems, synthetic content generators, and anonymized net-works. Through systematic analysis of 47 real-world cybercriminal cases and forensic evaluation against 12 at-tack vectors, SHARD achieves 89.2% attribution accuracy while reducing investigative timelines by 64% com-pared to conventional methods. We present novel techniques for blockchain temporal analysis, deepfake prove-nance tracking, and Tor-exit node correlation. The frame-work integrates machine learning-based anomaly detection with cryptographic verification to distinguish legitimate decentralized activity from adversarial manipulation. Our contributions include: (1) a formal threat model encompassing Web3 forensics; (2) a hybrid architecture combining on-chain and off-chain analysis; (3) algorithmic innovations for synthetic media fingerprinting; and (4) extensive empirical validation against contemporary attack scenarios. This work addresses a critical gap in digital forensics as investigative techniques must evolve alongside the technological infrastructure that criminals exploit.
The rapid proliferation of Internet of Things (IoT) devices across various industries, including healthcare, smart cities, and industrial automation, has introduced significant security, authenticity, and traceability challenges within increasingly complex supply chains. Although existing approaches have utilised blockchain-based digital identity solutions to address some of these concerns, persistent issues of counterfeit products and inadequate lifecycle transparency highlight the need for more robust, hardware-anchored identification mechanisms. Our work presents a novel architecture that integrates Physically Unclonable Functions (PUFs) and blockchain-based Soulbound Tokens (SBTs) to establish secure and verifiable digital identities directly tied to the physical hardware of IoT devices. By employing cryptographic tools such as fuzzy extractors, Merkle trees, and zero-knowledge proofs, the proposed architecture ensures accurate lifecycle tracking through key operational stages, including manufacturing, procurement, provisioning, maintenance, and eventual disposal or recycling. Performance evaluations conducted on the Ethereum Sepolia testnet demonstrate reasonable computational overhead in terms of gas usage and transaction confirmation times. The findings reveal that this approach aligns with NIST Special Publication 800-161 guidelines, as well as emerging regulatory standards, notably the European Union’s Digital Product Passport initiative, and has significant implications for enhancing transparency, sustainability, and security across global IoT supply chains.
Open access
Physical Unclonable Functions (PUFs) and Hardware Security
Audio piracy detection is increasingly complex in decentralised distribution settings, where mainstream approaches fail to ensure robustness, verifiability, or computational efficiency. Conventional Digital Rights Management (DRM) systems mainly enforce licensed access, but once content is copied or redistributed outside their control they offer little protection. Classical fingerprinting approaches such as MFCC based hashes can detect near-exact duplicates, yet they often fail under signal edits like pitch shifting, time stretching or equalisation. Deep learning embeddings improve robustness but demand heavy computation and centralised resources, making them less suitable for edge or decentralised deployments. These limitations call for a solution that is both edit resilient and verifiable. We propose HashWave, a blockchain-integrated perceptual hashing framework that combines robust audio fingerprinting with tamper-proof verification. The system fuses MFCC, chroma and chroma CENS, CQT, spectral contrast, and lightweight tempo/energy cues, applying operation-aware weighting via [Formula: see text] and constrained DTW for time-scale edits. Evaluated across GTZAN, FMA-A Dataset for Music Analysis, and MUSAN (SLR17) with over twenty signal-processing transformations, HashWave achieves AUC 0.957 and TPR@1%FPR 0.952, outperforming MFCC-only baselines and approaching deep embeddings at lower CPU cost. The blockchain layer, built on Ethereum and IPFS, ensures decentralised hash storage, duplication control, and verifiable authorship with average upload and contract execution times of 0.017 s and 0.044 s. Together, these results establish HashWave as a practical, scalable, and secure framework for piracy detection across streaming, podcasting, and Web3 ecosystems.
Open access
Advanced Steganography and Watermarking Techniques
Mr. DEVENDAR, Nandi J. Reddy, B.Sahasra, T.Srileka
Artificial intelligence and the quick development of photograph editing software in latest years have made it very simple to regulate virtual pix covertly. The authenticity and dependability of digital media utilized in social networks, journalism, and criminal proof have come below scrutiny because of manipulations like copy-circulate forgery and deepfake creation. The aim of this work is to perceive photograph forgeries via combining deep gaining knowledge of-based class techniques with traditional feature extraction methods.The cautioned device extracts precise neighborhood functions from input images the usage of the oriented speedy and turned around brief (ORB) algorithm. For powerful feature matching, 2-Nearest Neighbor (2NN) and Hierarchical Agglomerative Clustering (HAC) are then used. A Convolutional Neural community (CNN) model is trained to distinguish among authentic and manipulated photos by means of figuring out pixel-degree irregularities and texture changes if you want to growth type accuracy. examined on the publicly reachable MICC-F220 and MICC-F2000 datasets, the device outperforms baseline SVM strategies with a ninety% detection accuracy and a zero.1 false tremendous charge
Open access
Digital Media Forensic Detection
Generative Adversarial Networks and Image Synthesis
Non-fungible tokens (NFTs) have become a significant digital asset class, each uniquely representing virtual entities such as artworks. These tokens are stored in collections within smart contracts and are actively traded across platforms on Ethereum, Bitcoin, and Solana blockchains. The value of NFTs is closely tied to their distinctive characteristics that define rarity, leading to a growing interest in quantifying rarity within both industry and academia. While there are existing rarity meters for assessing NFT rarity, comparing them can be challenging without direct access to the underlying collection data. The Rating over all Rarities (ROAR) benchmark addresses this challenge by providing a standardized framework for evaluating NFT rarity. This paper explores a dimension reduction approach to rarity design, introducing new performance measures and meters, and evaluates them using the ROAR benchmark. Our contributions to the rarity meter design issue include developing an optimal rarity meter design using non-metric weighted multidimensional scaling, introducing Dissimilarity in Trades (DIT) as a performance measure inspired by dimension reduction techniques, and unveiling the non-interpretable rarity meter DIT, which demonstrates superior performance compared to existing methods.
Hasib Ahmed Md Khyrul Islam, Huy T. Vo, Aditya Rane
In the era of synthetic media, deepfake manipulations pose a significant threat to information integrity. To address this challenge, we propose TrustDefender, a two-stage framework comprising (i) a lightweight convolutional neural network (CNN) that detects deepfake imagery in real-time extended reality (XR) streams, and (ii) an integrated succinct zero-knowledge proof (ZKP) protocol that validates detection results without disclosing raw user data. Our design addresses both the computational constraints of XR platforms while adhering to the stringent privacy requirements in sensitive settings. Experimental evaluations on multiple benchmark deepfake datasets demonstrate that TrustDefender achieves 95.3% detection accuracy, coupled with efficient proof generation underpinned by rigorous cryptography, ensuring seamless integration with high-performance artificial intelligence (AI) systems. By fusing advanced computer vision models with provable security mechanisms, our work establishes a foundation for reliable AI in immersive and privacy-sensitive applications.
Open access
2 source records
Adversarial Robustness in Machine Learning
Digital Media Forensic Detection
Generative Adversarial Networks and Image Synthesis
Abhishek Bhattarai, Abdulhadi Sahin, Maryna Veksler, Ahmet Kurt · 7 authors
As cryptocurrencies have become increasingly used as an alternative to regular cash and credit card payments, the wallet solutions/apps that facilitate their use have also become increasingly popular. This has also intensified the involvement of these crypto wallet apps in criminal activities such as ransom requests, money laundering, and transactions on dark markets. From a digital forensics point of view, it is crucial to have tools and reliable approaches to detect these wallets on devices and extract their artifacts quickly with greater efficiency. However, with current research and trends, forensic investigators still need to manually extract these file artifacts, which delays the time-sensitive investigation findings. As mobile devices increasingly facilitate cryptocurrency transactions, there emerges a critical gap and need for automated evidence extraction to detect crucial artifacts preventing illicit activities. Therefore, in this paper, we present a comprehensive framework that incorporates various machine learning (ML), image processing, and natural language processing (NLP) approaches to enable fast and automated extraction/triage of crypto-related artifacts from Android and iOS devices. Specifically, our method can automatically detect which crypto wallet exists on the device, their artifacts (i.e., database/log files), along with the crypto-related images, web browsing data, and SMS conversations. For each type of data, we offer a specific ML technique, such as Support Vector Machine, Logistic Regression, and Neural Networks, to detect and classify these files. Our evaluation results show very high accuracy compared to alternative tools: our wallet classification model achieves 91% recall, crypto-related image classification achieves 75% accuracy, browsing data achieves 100% accuracy, and the SMS message model achieves 85% accuracy.
Abdullah Ayub Khan, Asif Ali Laghari, Syed Azeem Inam, Sajid Ullah · 6 authors
Rapid technological breakthroughs in recent years, like Deepfake, have made it feasible to produce synthetic media that is remarkably lifelike, but they also present significant hazards to public trust, privacy, and security. This survey paper reviews the latest techniques for detecting deepfakes, focussing on important components as image and video manipulation, audio spoofing, and multimodal synthesis. It features state-of-the-art methods including machine learning (ML), deep learning (DL), and multimodal architectures that are especially made to address the previously described deepfake criteria. The report provides a critical review of assessment measures used to assess detection model performance, including precision, accuracy, recall, computing effectiveness and efficiency, and fast responses to adversarial attacks. In order to assist direct future research, this highlights recent advancements in the subject, including explainable AI, federated learning, and self-supervised learning hierarchy. In order to examine the problems with adversarial attacks, scalability across different datasets, and the ethical implications of detection techniques, it is also vital to look into the technological and societal challenges surrounding multimedia-enabled deepfake detection. In particular, the usage of Blockchain Distributed Ledger Technology (BDLT) for traceability, lightweight modelling, and resilient systems forms for cross-model deepfake evaluation are discussed in this review study along with potential solutions to these limitations and areas for further research. This paper offers a comprehensive resource for future research, experts, and practitioners looking to combat the growing threat of deepfake, especially in the social media space, using innovative and useful detection tools.
Open access
Digital Media Forensic Detection
Generative Adversarial Networks and Image Synthesis
Atsuki Koyama, Kentaroh Toyoda, Manato Fujimoto, Thi Hong Tran
The rapid advancement of deepfake technology poses serious risks, including financial fraud and political misinformation, demanding robust methods for verifying image content authenticity. While the C2PA standard and zero-knowledgeproof-based methods provide an image content authenticity proving mechanism, the existing solutions struggle to efficiently support privacy-preserving edits and iterative modifications. To address these challenges, we propose zk-REAL (Zero-Knowledge-Based Protocol for Repeated Image Edit Authenticity Proof with Lattice Hashing), a framework that leverages a lightweight lattice-based hashing scheme within a zero-knowledge proof system. Our approach significantly reduces computational overhead, enabling faster proof generation and smaller proof size even for high-resolution images. Additionally, the updatability of our hashing method supports iterative edits, such as mosaicking or partial modifications, by minimizing redundant computations. Finally, to ensure compatibility with the C2PA ecosystem and conventional signature verifications, we integrate SHA-256 outside of the zero-knowledge circuit. Our evaluation shows up to a 29% reduction in computational costs for proof generation, showcasing the potential of zk-REAL in practical content authenticity verification scenarios.
Open access
2 source records
Advanced Steganography and Watermarking Techniques
Stefan Dziembowski, Shahriar Ebrahimi, Parisa Hassanizadeh
Ensuring the authenticity and credibility of daily media on internet is an ongoing problem. Meanwhile, genuinely captured images often require refinements before publication. Zero-knowledge proofs (ZKPs) offer a solution by verifying edited image without disclosing the original source. However, ZKPs typically come with high costs, particularly in terms of prover complexity and proof size. This paper presents VIMz, a framework for efficiently proving the authenticity of high-resolution images using folding-based zkSNARKs; a type of proving system that minimizes computational overhead by recursively folding multiple evaluations of the same constraints into a compact proof. As a complete proof system, VIMz proves the integrity of both the original and edited images, as well as the correctness of the transformation without revealing intermediate images within a chain of edits--only the final result is disclosed. Moreover, VIMz maintains the anonymity of the original signer and all subsequent editors while proving the authenticity of the final image. We also compare VIMz with the system model in Coalition for Content Provenance and Authenticity (C2PA) from different perspectives and show that VIMz offers higher level of security guarantee by eliminating the need to trust the editing environment. Experimental results show that VIMz performs efficiently in both prover and verifier sides. It can prove the transformations on 8K (33MP,i.e., 100MB) images with up to 13%~25% faster than the competition, while reaching to a peak memory of only 10 GB. Moreover, VIMz has a verification time of under 1 second and achieves succinct proofs of less than 11 KB for all resolutions, which is more than 90% improvement compared to the competition. VIMz's low memory complexity allows for proving multiple transformations in parallel to achieve a 3.5x additional speedup on average.
Open access
Advanced Steganography and Watermarking Techniques
Digital Media Forensic Detection
Physical Unclonable Functions (PUFs) and Hardware Security
Alanoud M. Almhlbdi, Norah D. Altowairqi, Areej Alshutayri, Rehab Qarout
Identifying hidden payloads in images has become increasingly critical as steganography continues to challenge traditional security measures. This paper introduces a deep learning framework for both the detection (binary classification) and fine-grained classification (multi-class) of steganographic payloads embedded using Least Significant Bit (LSB) techniques. The proposed system distinguishes between benign images and stego images containing five different payload types: HTML, JavaScript, PowerShell, URLs, and Ethereum-related data. To achieve this, we systematically evaluate various architectures, including a custom Convolutional Neural Network (CNN), hybrid CNN-GRU and CNN-LSTM models, and a Vision Transformer (ViT) at different input resolutions using 5-fold cross-validation. Our experiments reveal a critical finding: image resizing significantly degrades detection performance, as subtle LSB artifacts are often corrupted. While our custom CNN model achieved the highest mean cross-validation accuracy (0.9702), the hybrid CNN-GRU model demonstrated superior generalization on the held-out test set and external dataset, achieving a multi-class accuracy of 0.98 on the testset and 0.97 on the external. This result highlights the advantage of combining the CNN’s spatial feature extraction with the GRU’s ability to model sequential dependencies for robust payload identification on unseen data.
Open access
Advanced Steganography and Watermarking Techniques