The rapid proliferation of smart IoT devices has heightened the need for secure and efficient mechanisms to ensure the dependability of high-resolution image transmission while preserving computational performance. Current encryption methods often struggle to achieve a balance between robust security and resource efficiency, particularly in resource-constrained IoT environments such as smart cameras, mobile devices, and home automation systems. This paper introduces a novel lightweight image encryption algorithm designed to meet the stringent performance, security, and dependability requirements of IoT systems. The proposed algorithm incorporates a dynamic key expansion mechanism, a three-dimensional (3D) substitution box (S-box) structure, and a tightly integrated permutation process to enhance security and scalability. By dynamically adapting the key generation and encryption processes to image size, the scheme efficiently operates within a single encryption round while maintaining strong diffusion and confusion properties. The integrated permutation and substitution phases are optimized to enhance resilience against cryptographic attacks, such as differential and statistical attacks. Experimental evaluations on IoT devices validate the proposed algorithm's dependability and security, as demonstrated through rigorous correlation tests, entropy analyses, and sensitivity assessments. Comparative studies further establish the algorithm's superior performance in securing multimedia content while adhering to the real-time processing constraints of IoT systems. This work contributes a dependable and secure encryption framework for modern IoT applications, addressing the critical need for privacy and data integrity in increasingly connected environments.

Alzheimer’s disease (AD) is a serious neurodegenerative disorder worldwide with devastating consequences for cognitive function and quality of life. Early and accurate diagnosis of AD is important for effective clinical intervention and management. Existing artificial intelligence (AI) diagnostic models for AD are generally black-box systems, which diminishes their transparency and clinical acceptability. This issue is a major obstacle to integrating AI technologies into healthcare systems, where interpretability and clarity in decision-making are critical factors for gaining the trust of medical professionals. Based on these challenges, we propose an explainable ensemble framework for the early diagnosis of AD using structured clinical data from the AD Neuroimaging Initiative (ADNI). Our model integrates state-of-the-art ensemble classifiers-random forest, LightGBM, stacking, and voting ensembles-with SHAP (Shapley additive explanations) and fuzzy SHAP (FSHAP) to provide transparent feature attributions. We preprocess ADNI’s 180,799 records, which include 15 clinical and biomarker features, by handling missing values (through deletion or imputation), applying synthetic minority oversampling technique (SMOTE) for class balancing, encoding categorical variables, and correcting outliers. Among the evaluated models, the voting ensemble achieved the highest test accuracy ($98.6 \%$) along with robust precision and recall scores (approximately $98 \%$). SHAP analysis identified cognitive and neuropsychiatric scores, such as Mini-Mental State Examination (MMSE) Total, Functional Activities Questionnaire (FAQ), Neuropsychiatric Inventory Questionnaire (NPIQ), and Clinical Dementia Rating (CDR). FSHAP further quantified the confidence levels associated with each feature’s contribution, demonstrating that LightGBM models assign stronger and more definitive importance to certain features (e.g., Global CDR, FAQ), whereas ensemble averaging produces more conservative attributions under uncertainty.

As a key application within Internet of Things (IoT) engineering systems, secure communication protocols are essential for safeguarding sensitive information during data exchanges. Conventional and lightweight ciphers provide secure communication, but they often struggle to accommodate varying data sizes effectively while maintaining strong security. To address this limitation, the Tree-Feistel Cipher Standard (TCS) is proposed, a novel encryption scheme that integrates a hierarchical tree structure, an extended Feistel framework, and a substitution-permutation network (SPN). TCS employs a 512-bit block size and a 128-bit secret key, which undergoes adaptive operations to generate level-specific subkeys. A pre-processing phase employs a one-dimensional chaotic map to dynamically generate substitution-boxes (S-boxes) and permutation tables (P-tables), ensuring high nonlinearity and randomness. Encryption begins at level 3 with 64-bit blocks and leverages XOR operations, S-box lookups, and P-tables to provide strong confusion and diffusion. Experimental results demonstrate that TCS achieves superior performance compared to classical, lightweight, and state-of-the-art ciphers in terms of encryption speed, robustness, and security. TCS is especially suited for IoT and general communication scenarios, as validated through implementations on four platforms: a standard PC, a LiDAR sensor with Raspberry Pi, and two constrained devices (Teensy 3.6 and DOIT ESP32). Practical deployment within the OPC Unified Architecture (OPC UA) protocol further confirms TCS’s applicability in diverse computational environments.

Abstract Within the domain of image encryption, an intrinsic trade-off emerges between computational complexity and the integrity of data transmission security. Protecting digital images often requires extensive mathematical operations for robust security. However, this computational burden makes real-time applications unfeasible. The proposed research addresses this challenge by leveraging machine learning algorithms to optimize efficiency while maintaining high security. This methodology involves categorizing image pixel blocks into three classes: high-information, moderate-information, and low-information blocks using a support vector machine (SVM). Encryption is selectively applied to high and moderate information blocks, leaving low-information blocks untouched, significantly reducing computational time. To evaluate the proposed methodology, parameters like precision, recall, and F1-score are used for the machine learning component, and security is assessed using metrics like correlation, peak signal-to-noise ratio, mean square error, entropy, energy, and contrast. The results are exceptional, with accuracy, entropy, correlation, and energy values all at 97.4%, 7.9991, 0.0001, and 0.0153, respectively. Furthermore, this encryption scheme is highly efficient, completed in less than one second, as validated by a MATLAB tool. These findings emphasize the potential for efficient and secure image encryption, crucial for secure data transmission in rea-time applications.

As an integral application within the aerospace and electronic engineering systems, the aviation system relies heavily on secure communication protocols to safeguard sensitive information exchanged during travel and operations. However, many existing communication systems lack adequate security, leaving them vulnerable to adversarial attacks and data exfiltration. While effective in many security protocols, classical encryption algorithms are susceptible to various attacks, including statistical, differential, and side-channel attacks. To address these vulnerabilities, a new cryptographic encryption algorithm is proposed based on the extended Feistel network structure and binary tree structure. This algorithm enhances security measures by segmenting plaintext into blocks of 1024 bits, further divided into left and right halves across four distinct levels. Encryption begins at Level 64, employing a summation-based algorithm and xor operations to ensure diffusion and confusion properties. The parallel implementation of encryption enhances processing speed while deriving subkeys from a sensitive secret key enhances security against single-bit changes. Experimental assessments and security analyses demonstrate the robustness of the proposed cipher against various attacks. The proposed cipher offers high-quality encryption capabilities, making it an ideal candidate for inclusion in the secure communication protocols of aviation systems.

The escalating demand for accessible and efficient healthcare, coupled with advancements in artificial intelligence (AI), motivates a paradigm shift in medical consultations. This research introduces the Hayat Medical Consultation System, aiming to revolutionize communication and data management in healthcare. By integrating AI-driven solutions, the system addresses longstanding challenges, including secure medical image watermarking and accurate speaker diarization. This paper reviews existing literature on medical image watermarking and speaker diarization techniques, proposing an integrated solution that balances security, quality preservation, and robustness. The Hayat System’s approach encompasses secure image watermarking, advanced speaker diarization, AI-driven speech summarization, and a custom-built chatbot assistant. Tested on various aspects such as efficiency and security, the system effectively summarizes consultation descriptions using AI-based ChatGPT, allowing doctors to review and align summaries with diagnoses. Additionally, the system successfully encrypts data using RC4 and embeds it into medical images, addressing critical challenges in data transfer, integrity, and the secure management of patient records across different healthcare facilities.

Abstract In recent years, there has been a surge in new chaos-based cryptographic algorithms, many of which claim to have unusually large keyspaces. Although cryptographic primitives such as symmetric-key ciphers should have a secret keyspace large enough to resist brute force attacks, simply increasing the size of a secret key may not lead to improved security margins. An n -bit key may not necessarily have a keyspace of $$2^n-1$$ 2 n - 1 due to the key scheduling algorithm or how the key is used. In this paper, we cryptanalyse several chaos-based algorithms from the perspective of their key schedules. Our numerical analysis is based on the known-plaintext attack model, Kerckhoff’s principle and considers the number representations used for real number computation. Our analysis reveals that the actual security margins for these ciphers are significantly lower, some by a factor of over $$2^{100}$$ 2 100 than what was claimed. We then provide accurate keyspace estimates for these ciphers. Finally, we highlight alternative solutions for how secret keys can be used in the context of chaos-based cryptography and propose a simple key schedule as a proof of concept. Despite its simplicity, the proposed key schedule not only ensures that the keyspace matches the key length but also passes both the NIST and ENT statistical test suites, making it a viable option for generating secure cryptographic keys. Our work contributes towards addressing one of the fundamental problems in chaos-based cryptography that limits its real-world impact and reputation within the cryptographic community.

In the Industrial Internet of Things (IIoT), ensuring timely and secure data transmission between sensors and edge devices is paramount, particularly when dealing with sensitive information captured by high-resolution image sensors. However, existing methods often struggle to strike the delicate balance between security and efficiency, resulting in either vulnerable transmissions or significant processing delays. This article proposes a novel image encryption algorithm specifically designed for IIoT environments with constrained resources. The proposed algorithm leverages a new chaotic model that combines a cyclic construction with a 1-D perturbed logistic map. The proposed chaos model produces a single data sequence, subsequently utilized to generate three matrices mirroring the size of the image. Among these matrices, two contribute to crafting a permutation matrix for randomizing pixel positions, while the third matrix facilitates diffusion operations. The chaotic data sequence and generated matrices are preprocessed and can be used for encrypting multiple images under the same session key, enhancing efficiency. Encryption utilizes a single round that combines diffusion and permutation operations simultaneously, further reducing processing time. Experimental results demonstrate its effectiveness on an IoT camera sensor for encryption and a separate device for decryption. Statistical tests confirm the robustness of the encrypted images against various attacks, including correlation analysis, histogram analysis, differential attacks, and key and plaintext sensitivity. Furthermore, comparisons with existing image encryption techniques showcase the proposed algorithm's superior security and efficiency. Notably, it effectively encrypts images of varying sizes, making it suitable for deployment in IIoT environments.

Voice over Internet Protocol (VoIP) networks have recently emerged as a promising telecommunication medium for transmitting voice signal. One of the essential aspects that interests researchers is how to estimate the quality of transmitted voice over VoIP for several purposes such as design and technical issues. Two methodologies are used to evaluate the voice, which are subjective and objective methods. In this study, the authors propose a non‐intrusive machine learning‐based (NIML) objective method to estimate the quality of voice. In particular, they build a training set of parameters – from the network and the voice itself – along with the quality of voices as labels. The voice quality is estimated using the perceptual evaluation of speech quality (PESQ) method as an intrusive algorithm. Then, the authors use a set of classifiers to build models for estimating the quality of the transmitted voice from the training set. The experimental results show that the classifier models have a valuable performance where Random Forest model has superior results compared to other models of precision 94.1%, recall 94.2%, and receiver operating characteristic area 99.2% as evaluation metrics.