Biometric authentication provides high convenience with the drawback of privacy leakage, replay attacks, and centralized control over biometric templates. This paper introduces an Ethereum-based decentralized biometric authentication framework that uses Elliptic Curve Digital Signature Algorithm (ECDSA), InterPlanetary File System (IPFS) storage, and an on-chain challenge–response protocol. In the proposed model, encrypted biometric templates are stored of-chain in IPFS, whereas their content identifiers (CIDs) are registered in Ethereum smart contracts. Every authentication attempt necessitates a new on-chain nonce and an ECDSA signature of the concatenation of the CID and the nonce, authenticated through Ethereum’s built-in method, ecrecover. The design supports explicit replay protection, revocation, and public auditability. Deployment of the prototype on Ganache and MetaMask reveals that the scheme provides secure, transparent, and tamper-proof authentication with minimal gas consumption on FVC2004 datasets and reasonable storage usage on Ethereum.
Open access
Cryptography and Data Security
Advanced Steganography and Watermarking Techniques
Кваліфікаційна робота бакалавра складається зі вступу, чотирьох розділів, висновків, списку використаних джерел та додатків. У першому розділі виконано аналіз предметної області, наведено приклади існуючих рішень та визначено основні функціональні потреби користувачів. Другий розділ містить специфікацію функціональних і нефункціональних вимог до інформаційної системи, її логічну структуру та ключові компоненти. У третьому розділі було здійснено розробку вебзастосунку з підтримкою смарт-контрактів для безпечного виконання криптотранзакцій на базі бібліотеки React та мови програмування Solidity, доповнивши цей технологічний стек хмарною платформою Supabase та бібліотекою Ethers.js для взаємодії з криптогаманцем MetaMask. У четвертому розділі наведено спеціальні розрахунки, зокрема здійснено ергономічну оцінку функціональних компонентів програмного продукту з використанням математичних моделей, розраховано час на виконання повного циклу замовлення в Web3-системі з урахуванням мережевих затримок та фізичного введення даних, а також обґрунтовано високу ефективність спроєктованого інтерфейсу. У висновках узагальнено результати виконаної роботи та визначено перспективи подальшого розвитку системи.
The fast digitalization of contemporary society has changed the data into a valuable resource, and it has been the key to the innovation in the financial sector, healthcare, politics, and industries, and it has also increased risks both in terms of misusing it, stealing it, and using it. Information security through maintaining confidentiality, integrity and availability of information has thus become a pre-requisite to trust in digital infrastructures. The present paper gives a detailed discussion of how cryptography, cybersecurity, and data privacy come into convergence and have a central role to play in protecting the digital ecosystems. Basic cryptographic primitives such as symmetric and asymmetric encryption, hash functions as well as digital signatures are discussed as the foundation of secure communication. With these, more complex privacy-sensitive technology like homomorphic encryption, zero-knowledge proofs, and differential privacy is discussed as technology that could offer the opportunity to perform safe computation and share data without jeopardizing the privacy of individuals.The paper also explores the disruptive potential of quantum computing, specifically how it can render the popular public-key systems insecure by figuring out ways to break them, e.g. the Shor algorithm, and assesses the new paradigm of post-quantum cryptography as a reaction to this existential risk. The examples are discussed within various fields such as secure communication schemes, data-at-rest security, cloud computing, and the Internet of things (IoT), e.g., in which cryptographic efficiency and versatility are most crucial. It is a synthesis of these factors that the paper highlights that cryptography is not only a technical protection but it is a cornerstone enabling resiliency, trust, and privacy-by-design in the digital era. This paper then ends with a discussion on the challenges that still need to be tackled, including scalability, usability and compliance with regulations, and how future research will be needed to define the future of secure and privacy-preserving technologies in the increasingly interconnected world.
To address centralized trust risks, inadequate privacy protection, and quantum vulnerability of traditional crossdomain authentication systems, this paper proposes a quantumresistant self-sovereign identity (SSI) scheme based on lattice cryptography. Centered on Decentralized Identifiers (DIDs) and Verifiable Credentials (VCs), it constructs a decentralized ecosystem with two key innovations: a “single trapdoor - multi-attribute public key” structure (reducing storage and simplifying key management) and a lattice-based linkable ring signature (balancing anonymity and traceability). Implemented via TrapGen, SamplePre, and rejection sampling, the scheme's security relies on the Short Integer Solution (SIS) problem, with unconditional anonymity and unforgeability proven in the random oracle model. Efficiency comparisons confirm advantages in signature/verification time and storage overhead. This work provides a secure, privacy-preserving post-quantum solution for crossdomain collaboration in smart cities and supply chains. Future work will focus on lattice optimization and zero-knowledge proof integration.
Maiesha Fahomida, Nushraq Nawer Hossain, Farhan Ahmad Nafis, Raian Islam
Modern digital communication requires stronger mechanisms for both confidentiality and authentication to mitigate threats such as impersonation, replay, and eavesdropping. Although traditional cryptographic methods offer secrecy, they lack strong identity verification in adversarial environments. To ensure security in both transmission and authentication, we proposed a hybrid framework combining the most effective mechanisms for secure communication, enhanced with LSB Steganography to conceal sensitive information. Zero-knowledge proofs are used for secure authentication. Diffie-Hellman with AES-256-GCM ensures confidentiality and data integrity, while LSB Steganography provides secure concealment of transmitted communication. Our method has been evaluated using a variety of techniques, including steganographic quality assessment, encryption-decryption performance testing, and authentication time measurement, confirming its resilience against common security risks. The proposed methods achieve PSNR values up to$\mathbf{7 4. 5 8 ~ d B}$and SSIM of 0.9999. The encryption time ranges from 0.035 ms to 0.068 ms, while the decryption time remains consistently lower, ranging from 0.008 ms to 0.015 ms. The results demonstrate that the proposed framework is a viable option for secure data transfer, as it guarantees confidentiality, integrity, authentication, and covert communication.
The article presents an analysis of the robustness of an authentication scheme based on zero watermarking. The study examines a two-factor authentication scheme that uses "knowledge of something" (a password) and "possession of something" (a digital RGB image) as its factors. The zero watermarking algorithm chosen is based on DWT and K-means transformations, with additional use of the Swish function. The analysis is conducted by considering the theoretical complexity of the algorithm assuming the adversary knows its parameters, such as the password, the hash of the password, the image, the reference watermark, the transformation result, and other parameters. Previous studies have shown a high theoretical robustness of the scheme, which relies on the complexity of the password and the dimensionality of the image. For large image sizes (512×512 pixels and above), a relatively high level of cryptographic resistance is achieved. However, this robustness is not formally proven, and the actual strength may be significantly lower due to the specifics of the images and transformations, which can introduce additional vulnerabilities. The algorithm is subject to a relatively high rate of collision, associated with digital image transformations and matrix multiplications, which weakens its resistance. Authentication schemes and zero watermarking algorithms require further research, formal proof of cryptographic properties, and methods for integration into access control systems, as they can provide a high level of authentication robustness in systems with high noise levels. Additionally, the convenience and low cost of such schemes give them an advantage over other authentication methods. The study provides recommendations for improving the potential characteristics of the algorithm.
Open access
Advanced Steganography and Watermarking Techniques
The growth of cloud computing in the healthcare field has led to significant developments, but ensuring the confidentiality and protection of medical records such as electronic health records (EHRs) remains a major concern for healthcare service applications. In cloud computing, the basic authentication provided by most service providers is insufficient to ensure secure access to critical or sensitive resources. Moreover, most of the existing healthcare management systems are ineffective in handling a number of patient data, which leads to single points of failure. To address these issues, elliptic curve cryptography (ECC) with Curve25519 is utilized to enhance security in cloud storage, particularly within healthcare management systems. The ECC with Curve25519 is optimized for efficient and fast scalar multiplication, which reduces computational overhead and enhances performance. The curve parameters are selected to prevent vulnerabilities and ensure security against known attacks. Moreover, it is efficient in maintaining the integrity of patient records, which reduces storage and bandwidth requirements. The ECC with Curve25519 achieves lower Key-Gen, prove, verify, proving key size, and verification key size of 13.7 s, 48 s, 0.608 s, 13.27 Mb, and 123.70 Kb, respectively, in comparison with proxy re-encryption algorithm with zero-knowledge proof (ZKP).
The paper explores the possibility of expanding the use of end-to-end encryption protocols based on the Double Ratchet algorithm in applications with low trust in the server, particularly in turn-based games and strategic interactions. The relevance of the research is due to the growing need for secure communication in cyberattacks, especially during military operations. The field of end-to-end encryption requires the study of additional applications beyond the usual ones, such as encrypted communication in text messengers. The developed implementation of the protocol can be safely used in any applications that aim to implement end-to-end encryption and satisfy the criterion of session ephemerality (in cases where secrets are stored outside a secure environment). The implemented server supports ephemeral sessions, which guarantee minimal risks of information compromise, and uses digital signatures (EdDSA) for user authentication. Logical routing of requests ensures efficient message transmission in secure scenarios. The choice of the classic game of checkers as an example allowed the authors to effectively demonstrate the advantages of end-to-end encryption and the capabilities of the implemented protocol. All cryptographic operations, including key generation, encryption and decryption of messages, are successfully performed on client devices. It is important to improve error handling mechanisms and optimize the operation of WebAssembly. An interesting area of further research is the creation of zero-knowledge proof mechanisms to prevent Man-In-The-Middle attacks during the creation of a shared secret, optimizing integration with cryptographic hardware security modules (HSM), and exploring the scalability of the solution. The proposed approach can be used to solve real-world information security problems where trust in the data transmission channel is critically important. Thus, the work has created a comprehensive solution that includes a cryptographic protocol, a backend, and a web client, which demonstrates the viability of end-to-end encryption in browser environments and multiplayer games. The work can be used as a basis for further research and development in the field of security of communication systems and privacy in multiplayer games.
10.5281/zenodo.17605813 chaos structure complexity sequences / test files / public domain Chaos Complexity Domain Sequencing"Maximum Entropy Equilibrium"sha384sum OUTFN_BASE-OUTFN_VER-OUTFN_VERMIN-20221230191340.OUTFN_EXT.1069cbf8cebedf73040848960d915d728f8ebce64de339e57c03984b9b125065571ee73cba2fbe8324d57770631f22d3c27download Value Char Occurrences Fraction 0 4000000106 0.500000 1 3999999894 0.500000Total: 8000000000 1.000000Entropy = 1.000000 bits per bit.Optimum compression would reduce the sizeof this 8000000000 bit file by 0 percent.Chi square distribution for 8000000000 samples is 0.00, and randomlywould exceed this value 99.81 percent of the times.Arithmetic mean value of data bits is 0.5000 (0.5 = random).Monte Carlo value for Pi is 3.141394237 (error 0.01 percent).Serial correlation coefficient is 0.000013 (totally uncorrelated = 0.0).sha384sum OUTFN_BASE-OUTFN_VER-OUTFN_VERMIN-20230103155948.OUTFN_EXT.107d6275873f72a0edc7585db17bba50cdfb4a5097f3f50ac7b69eaa85ae8ec95975fb187579a5b05ff0c69ca71378fe71d download Value Char Occurrences Fraction 0 4000000107 0.500000 1 3999999893 0.500000Total: 8000000000 1.000000Entropy = 1.000000 bits per bit.Optimum compression would reduce the sizeof this 8000000000 bit file by 0 percent.Chi square distribution for 8000000000 samples is 0.00, and randomlywould exceed this value 99.81 percent of the times.Arithmetic mean value of data bits is 0.5000 (0.5 = random).Monte Carlo value for Pi is 3.141257485 (error 0.01 percent).Serial correlation coefficient is 0.000004 (totally uncorrelated = 0.0). DATA MORGANA COMMUNICATIONS AUTHOR/ EDWIN J. VENINGEDITOR EDWIN J. VENINGCORRESPONDENCE ADMIN@DATAMORGANA.NETWEBSITE SPAWN HTTPS://WWW.DATAMORGANA.NETRELEASED DD 20230525 [ YYYYMMDD ]EDIT REV.DD 20231208 over 20230726REF <symbolic base> See addendum:- binary ambiguity is expectedIntroduction in Dutch : page 2 20230726 crt0 Addendum: The test vectors presented here stem from the design of a custom generator, originally intended to outperform competitors in various categories of "randomness" generation. The goal was to achieve chaotic streams that exceeded the capabilities of other contenders, without relying on traditional methods for balancing distribution qualities. The resulting system incorporates parametric high-gain, maximum entropy equilibrium functions and methods, with output files available for download from this page. These files are derived from this work and should be used with caution. Historical Context: In 2019, a proposal was made to enhance the cryptographic subsystem of operating systems through a novel approach. This concept involved hardening the system with a new cryptographic processing "idea" of operation(s), integrated within a fresh confidence model. This idea was presented as the open-source project: /dev/entropy, a Unix non-blocking character device designed for non-disclosed ZKP (Zero-Knowledge Proof) seasonal or projected transactions/operations. The goal was to bootstrap system entropy pools using unique host identification, confidence constraints, and host signature processing in its own ZKP design (a system verifier capsule). /dev/entropy was intended to serve as the system entropy pool, which would be well-documented and securely stored. The author and programmer asserted that chaining cryptographic functions could weaken their security, leading to a proposal for entropy pools that would re-seed cryptographic functions in the host stack using non-linear, complexity-driven methods. These operations were intentionally designed to be opaque to prevent exposure, aiming to mitigate known mechanical noise attack vectors and thwart binary dissection. The processing would involve a novel use of "RAM" or "held latent memory." The project concluded in 2019 but remains a significant influence on the development of unique event processing and symbolic information transformations. As for the test vectors, no claims are made regarding their randomness or indexing properties. Envisioned Applications for the Methods and Functions: High-speed calibration of scientific instruments High-gain precision, offering persistent increases in resolution for guidance systems, telemetry, and high-availability scheduling (real-time systems) Persistence of identification tokens, tokenizing information by range, sequence hinting (*), as suggested in the ZKP paper ZKP 'circuitry' / 'gadgets' with enhanced properties, allowing for directional confidence balancing and omni-directional jumps, encoding with unique event processing such as spacetime locality encoding Real-time processing improvements, introducing new priority-type scheduling and domain sequencing (correlated context, with no known limits or recursion results) Application of "lossy" parity and "hashing" in new contexts, utilizing range hinting or the development of a symbolic encoded sequence that persists in noisy systems. The ratio is under testing. Expected hardware development: Domain sequencing through event processors with hardened/optical circuitry and one-way functions These methods aim to serve as a critical infrastructure carrier post-quantum Cryptography (PQC), offering potential solutions for complex network topologies and signal semantics for future interstellar applications. This approach leverages spatial and referential qualities without sudden collapse, adding the Temporal Domain Cryptography from 2015 as part of the ongoing evolution. DISCLAIMER: The contents of these vectors may contain the densest information to date, with an inherent carbon footprint that requires careful handling. Due to the dense nature of this data, it may cause local mechanical friction and, in extreme cases, could lead to combustion. As with any significant discovery, proceed with caution. Note: This is not the recommended practice in the narrowing binary domain of information. For reference: CACert Random Number Results — "No Entropy Here" home https://www.datamorgana.net
Xin Liu, Anyang Qi, Lanying Liang, Dan Luo · 10 authors
In computer vision, the intersection determination of polygonal areas is utilized to segment different regions in an image and assist in detecting the boundaries of the regions. Moreover, the secure computation of the intersection area of polygons can solve the private calculation of geometric problems in machine learning. A security protocol under the semi-honest model was designed for the problem of secure computation of the intersection area of two polygons. This protocol adopts a new coding method and the Paillier homomorphic encryption algorithm. Aiming at the malicious behaviors that malicious participants may carry out in the semi-honest protocol, a secure computation protocol for the intersection area of polygons under the malicious model was designed by using methods such as hash function, cut-and-choose and zero-knowledge proof. The security of this protocol was proved, and its computational complexity and communication complexity were analyzed. Compared with the existing schemes, it is more efficient.
Random numbers are basic building blocks for cryptography. For example, they are heavily utilized in Decentralized Finance (DeFi) and blockchain applications. Cryptographers and practitioners frequently employ bit selection, arithmetic, and logical operations to generate cryptographically secure random numbers (CSPRNs), thereby achieving the desired level of entropy and security. There is a need to analyze the security of such operations on CSPRNs. In this paper, we have studied and analyzed the security properties of arithmetic and some string operations on CSPRNs, and reviewed Boolean logic operations with a focus on the preservation or loss of entropy. We have analyzed and presented several proofs of security or lack of it for such operations. We have implemented and conducted experiments to corroborate these results using the NIST test suite. Our work applies not only to classical random numbers but also to quantum random numbers.
The convergence of quantum physics and machine learning presents unprecedented opportunities for developing ultra-secure authentication systems. This comprehensive paper investigates the integration of quantum random number generators (QRNGs) with advanced machine learning architectures, including quantum neural networks (QNNs), long short-term memory (LSTM) networks, and hybrid quantumclassical models, to establish authentication mechanisms with information-theoretic security guarantees. We provide rigorous theoretical foundations spanning quantum entropy theory, min-entropy estimation, and randomness certification, complemented by detailed analyses of contemporary QRNG hardware implementations including photonic integrated circuits achieving generation rates exceeding 20 Gbps. The paper explores deep learning architectures for biometric authentication, demonstrating how QNN-enhanced systems achieve superior performance through quantum superposition and entanglement. Furthermore, we examine the application of quantum entropy sources in zero-knowledge proof protocols, particularly zk-SNARKs and zk-STARKs, addressing post-quantum security concerns. Through comprehensive mathematical formulations, algorithmic implementations, and security analyses, we establish that hybrid quantum-classical authentication systems combining QRNG-derived cryptographic keys with ML-based behavioral authentication provide provably secure, practical solutions for next-generation cybersecurity applications. Experimental results from current quantum hardware platforms validate theoretical predictions and demonstrate real-world applicability.
Open access
Chaos-based Image/Signal Encryption
Physical Unclonable Functions (PUFs) and Hardware Security
Abstract — The Fractal Eavesdrop Detection (FED) protocol defines a cryptographic mutual-authentication... The Fractal Eavesdrop Detection (FED) protocol defines a cryptographic mutual-authentication and integrity validation mechanism between two fractal nodes sharing a recursive lineage. Unlike conventional systems that rely on fixed keys or static hashes, FED uses algorithmic mutability, session-based seed derivation, multi-point challenge validation, and time-bound CRC binding to detect both impersonation and passive eavesdropping. The protocol is designed for lightweight, low-power devices such as ESP32-class microcontrollers and operates without blockchain consensus or zero-knowledge proofs, while still enabling secure proof-of-origin and tamper-awareness. FED serves as the security and validation layer within the EQUORA Institute’s Fractal Economy architecture and complements the BlockFractal cryptographic tokenization layer and the EquoraVault hardware-based proof-of-impact system. This document is released as part of the EQUORA Institute White Paper Series and is a preprint version (v0.8), subject to revision. All versions remain archived for DOI-based citation integrity.
Open access
2 source records
Chaos-based Image/Signal Encryption
Physical Unclonable Functions (PUFs) and Hardware Security
Blockchain technology into the healthcare sector, logistics, finance, and massive IoT system, this technology is making its way into it. With the issue of privacy becoming much more critical than before in the adoption phases. Even though blockchains provide integrity and resistant storage against tampering, numerous implementations still expose behavioural evidence, time-based data, besides other forms of metadata, which may be compromised by inference attacks. These breaches become more apparent when there is an exchange or near real-time processing of sensitive information. This paper will analyze the performance of three widely discussed privacy-saving methods Zero-Knowledge Proofs (ZKP), Homomorphic Encryption (HE), not to mention Differential Privacy (DP) in a practical blockchain architecture and not in isolation. A testbed made of Hyperledger Fabric and Ethereum instances privately owned was constructed to enable each of the methods to be tested under the same conditions. The experiments made trade-offs of transaction suspensions, added computational load, and exposure to metadata-based attacks.The results demonstrate that ZKP-based settings store observable metadata substantially and restricted the amount that an adversary would deduce regarding transaction patterns. HE was a good implementation in areas where safe computation was needed but with significant latency overhead. The least significant communication impact was on DP, which was, however, to be carefully tuned in terms of its privacy parameters to prevent the decline in protection. We further conducted simulated attacks to research into the reaction of each approach to pressure and detail how validation can be extended in the future with the aid of open datasets through the MIMIC-III, OpenAQ, and TII cybersecurity logs. On the whole, the findings provide useful insights to teams, who want to embed cryptographic privacy tools in production-scale blockchain systems in particular in those industries where confidentiality and data protection cannot be jeopardized.
Non-Fungible Tokens (NFTs) have gained attention as a technology for guaranteeing ownership of digital content, leading to rapid market expansion. However, NFTs are limited in that they guarantee ownership only for a single, explicitly designated digital asset. For instance, if an image associated with an NFT undergoes modifications such as resolution reduction or trimming, it falls outside the scope of the NFT’s guarantee. In this study, we propose a new NFT scheme capable of guaranteeing ownership for multiple digital assets that fall within a defined visual similarity threshold. The core of this method lies in replacing conventional cryptographic hash functions with Image Hash functions, allowing the scope of ownership to cover a "range" of similar content rather than a single exact match. This enables highly similar content to be automatically included within the NFT’s scope of guarantee without explicit designation. To verify the feasibility of this scheme, we implemented and evaluated a prototype using four types of Image Hash functions against common image transformations, such as resolution reduction and trimming, on the Polygon blockchain. The results indicate that both Average Hash (aHash) and Perceptual Hash (pHash) are suitable functions, and that the NFT verification process can be performed efficiently. This method provides a novel mechanism to dynamically extend the scope of NFT ownership, paving the way for new potential applications for NFTs.
Advanced Steganography and Watermarking Techniques
Rini Hardiyanti, Amil Ahmad Ilham, Ady Wahyudi Paundu
This research proposes the use of AES-256-CBC encryption, decentralized storage using Interplanetary File System (IPFS), and CID hash logging into the Ethereum. The system was tested using files of 5, 10, 15 and 100 MB, and shows that the encryption-decryption process has no significant impact on system performance. The security evaluation resulted 6 out of 7 test scenarios successfully prevented data theft, manipulation, and duplication after using attack simulations such as reply attacks, bit-flipping, Man-in-the-Middle (MITM), and Brute Force attacks. The total cost for logging CID hashes to Ethereum was $0.00913-0.01080$ ETH. Although throughput performance and execution time are volatile in both proposed and comparison system, the proposed system is superior in terms of system security and cost efficiency, making it feasible to use for e-learning content protection.
Zero-Knowledge Proofs (ZKP) and Homomorphic Encryption (HE) are crucial for data privacy in applications like cloud, blockchain, and analytics. However, the real-world adoption often faces performance challenges, particularly in the execution of the Number Theoretic Transform (NTT) required for polynomial multiplication involving sizes beyond \(2^{20}\) and large integer widths (e.g., 256 bits). FPGAs offer a promising platform for acceleration, but efficiently implementing large-size NTTs remains difficult due to the limited on-chip resources. The widely adopted four-step NTT method, used to relieve the need for large on-chip memory, introduces performance bottlenecks. Initially, the traditional dataflow NTT architecture may not fully exploit available compute capability, which hinders achieving peak performance. Furthermore, during the matrix transpose phase, the non-sequential access to external High-Bandwidth Memory (HBM) causes inefficiency. To address these challenges, we introduce HiFA, an FPGA-based automatic accelerator framework designed for high-performance and flexible large-size NTT computations. HiFA utilizes a stacked NTT architecture for high parallelism, maximizing HBM throughput. It supports various decomposed polynomial sizes via a novel reordering module. Additionally, a specialized cyclic shuffle module is integrated to optimize data movement during the matrix transpose step, alleviating random memory access delay. HiFA also provides an automatic Design Space Exploration (DSE) framework that identifies optimal four-step decomposition parameters and generates corresponding hardware configurations. Our experiments show that the FPGA implementation of HiFA achieves an average speedup of 2.97× and up to 7.25× improvement in latency over prior state-of-the-art FPGA solutions. Compared to prior GPU-based methods, HiFA achieves an average energy efficiency gain of 2.24×.
A. G. Ramakrishnan, Shubham Agarwal, Sharmila Kumari Selvanayagam, Kunwar P. Singh
As image generation models grow increasingly powerful and accessible, concerns around authenticity, ownership, and misuse of synthetic media have become critical. The ability to generate lifelike images indistinguishable from real ones introduces risks such as misinformation, deepfakes, and intellectual property violations. Traditional watermarking methods either degrade image quality, are easily removed, or require access to confidential model internals – making them unsuitable for secure and scalable deployment. We are the first to introduce ZK-WAGON, a novel system for watermarking image generation models using the Zero-Knowledge Succinct Non-Interactive Argument of Knowledge (ZK-SNARKs). Our approach enables verifiable proof of origin without exposing model weights, generation prompts, or any sensitive internal information. We propose Selective Layer ZK-Circuit Creation (SL-ZKCC), a method to selectively convert key layers of an image generation model into a circuit, reducing proof generation time significantly. Generated ZK-SNARK proofs are imperceptibly embedded into a generated image via Least Significant Bit (LSB) steganography. We demonstrate this system on both GAN and Diffusion models, providing a secure, model-agnostic pipeline for trustworthy AI image generation.
Open access
2 source records
Physical Unclonable Functions (PUFs) and Hardware Security
Adversarial Robustness in Machine Learning
Generative Adversarial Networks and Image Synthesis
ABSTRACT This study investigates an audit data privacy protection mechanism based on blockchain technology and constructs a secure and efficient computational model. The system is designed to support practical domains such as medical record systems and financial audit platforms, ensuring data integrity, traceability, and confidentiality. Leveraging a distributed ledger and optimized consensus mechanism, the model automates data sharing and audit processes through smart contracts. A hybrid encryption approach is proposed, integrating RSA algorithm with chaos theory to enhance encryption complexity and randomness. Experimental results—conducted on real‐world medical audit data—demonstrate that, compared to baseline methods, the proposed scheme improves privacy protection by up to 50%, increases ciphertext complexity by 45%, and reduces encryption and verification time by approximately 20 s. The system also supports up to 500 concurrent users with a throughput of 583.49 requests/s, indicating strong scalability and efficiency.
Advanced Steganography and Watermarking Techniques
Conventional cryptographic approaches face increasing insecurity because quantum computing and attacks enabled by AI operate at a rapid speed of technological advancement. The review analyzes IEEE research activities on progressive cryptographic methods consisting of post-quantum cryptography (PQC) and fully homomorphic encryption (FHE) in addition to zero-knowledge proofs (ZKP) and AI-augmented cryptographic models. We explore both advantages and limitations in addition to applicable uses for protecting modern digital frameworks which include blockchain, IoT and 5G networks. Standardization efforts as well as hybrid crypto framework trends receive analysis in the article to establish long-lasting secure systems.