Computers & Electrical Engineering最新文献

{"title":"Two-layer intelligent planning model of active distribution network based on simulated annealing algorithm and multi-objective function","authors":"He Huang ,&nbsp;Tao Lin ,&nbsp;XiaoDi Zhang ,&nbsp;Liang Liang ,&nbsp;JiYun Liu","doi":"10.1016/j.compeleceng.2025.110733","DOIUrl":"10.1016/j.compeleceng.2025.110733","url":null,"abstract":"<div><div>Traditional power distribution network planning methods face challenges such as single-objective optimization and the use of algorithms that are prone to becoming trapped in local optima, making it difficult to satisfy the multi-dimensional and complex requirements of modern distribution systems. To address these issues, this paper proposes a two-layer intelligent planning model based on simulated annealing and multi-objective functions, designed for active distribution networks (ADNs). The upper layer aims to minimize the total cost of the ADN, while the lower layer focuses on reducing voltage deviation and maintaining power system stability, leveraging the global search capability of the simulated annealing algorithm. Experimental results demonstrate that the proposed model reduces total costs by approximately 15 % to 30.8 %, with an average cost reduction per node of about 27 %, while effectively maintaining voltage deviations within the range of 0.98 to 1.03 per unit. The model successfully overcomes the limitations of single-objective optimization and poor algorithmic convergence in active distribution network/system (ADN/ADS) planning, exhibiting excellent performance in both cost efficiency and voltage stability enhancement, and offering an innovative solution for large-scale ADS planning.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110733"},"PeriodicalIF":4.9,"publicationDate":"2025-10-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266183","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Zero day malware detection with Alpha: Fast DBI with Transformer models for real world application","authors":"Matthew Gaber,&nbsp;Mohiuddin Ahmed,&nbsp;Helge Janicke","doi":"10.1016/j.compeleceng.2025.110751","DOIUrl":"10.1016/j.compeleceng.2025.110751","url":null,"abstract":"<div><div>The effectiveness of an AI model in accurately classifying novel malware hinges on the quality of the features it is trained on, which in turn depends on the effectiveness of the analysis tool used. Peekaboo, a Dynamic Binary Instrumentation (DBI) tool, defeats malware evasion techniques to capture authentic behavior at the Assembly (ASM) instruction level. This behavior exhibits patterns consistent with Zipf’s law, a distribution commonly seen in natural languages, making Transformer models particularly effective for binary classification tasks.</div><div>We introduce Alpha, a framework for zero-day malware detection that leverages Transformer models, Support Vector Machines (SVMs) and ASM language features. Alpha is trained on malware and benign software data extracted at the ASM level, enabling it to detect entirely new malware samples with exceptional accuracy. Alpha eliminates any common functions from the test samples that are in the training dataset. This forces the model to rely on contextual patterns and novel ASM instruction combinations to detect malicious behavior, rather than memorizing familiar features. By combining the strengths of DBI, ASM analysis, and Transformer architectures, Alpha offers a powerful approach to proactively addressing the evolving threat of malware. Alpha demonstrates excellent accuracy for Ransomware, Worms and APTs with flawless classification for both malicious and benign samples. The results highlight the model’s exceptional performance in detecting truly new malware samples.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110751"},"PeriodicalIF":4.9,"publicationDate":"2025-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145267313","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Two way amplify and forward relay with energy harvesting: Performance analysis","authors":"Kaustubh Ranjan Singh,&nbsp;Parul Garg","doi":"10.1016/j.compeleceng.2025.110762","DOIUrl":"10.1016/j.compeleceng.2025.110762","url":null,"abstract":"<div><div>In this work, we consider a wireless communication network between two node terminals <span><math><mrow><mi>A</mi><mo>,</mo><mi>B</mi></mrow></math></span> which is aided by a two way relay <span><math><mi>R</mi></math></span>. Besides the relay path, we assume an additional direct path between <span><math><mrow><mi>A</mi><mo>,</mo><mi>B</mi></mrow></math></span>. The relay <span><math><mi>R</mi></math></span> is energy constrained and thus performs energy harvesting (EH) by Time Switching (TS) architecture to meet its energy requirements. The relay forwards the received signal through Amplify and Forward (AF) protocol to the terminals which then perform selection combining (SC) for the signals received through direct and relay paths and decode the transmitted message. Expressions for system outage probability with asymptotic analysis, ergodic capacity are derived in closed form under Nakagami-m fading conditions. Further, the impact of parameters like TS factor are studied on the outage probability, ergodic capacity, energy efficiency and system throughput. The results obtained are verified through Monte Carlo simulations.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110762"},"PeriodicalIF":4.9,"publicationDate":"2025-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145267248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Low-latency and interpretable intrusion detection for IIoT using self-supervised learning with entropy-based masking","authors":"Fasih Ullah Khan,&nbsp;Adnan Noor Mian","doi":"10.1016/j.compeleceng.2025.110753","DOIUrl":"10.1016/j.compeleceng.2025.110753","url":null,"abstract":"<div><div>The Industrial Internet of Things (IIoT) has enhanced data connectivity across domains like smart city and industry. But this advancement has also created several security risks necessitating robust security measures. One critical challenge in developing effective intrusion detection systems (IDS) for IIoT is class imbalance in training datasets. In most cases, benign traffic predominates, leading to biased model training and underperformance in detecting rare attacks. To address these issues and effectively detect both normal and various attack categories, even with label scarcity and class imbalance, we propose a low-latency gradient boosting framework for efficient intrusion detection. Our approach uses Self-supervised learning (SSL) to improve efficiency and robustness. This hybrid approach employs a Masked Autoencoder (MAE) for robust representation extraction from unlabeled data, followed by classification using LightGBM. To enhance the learning capability of proposed framework, we fuse an entropy-based masking strategy within the MAE. This allows features with high uncertainty to be masked with high probability during training. This targeted feature selection enables the model to reconstruct the most informative features. As a result, the model’s robustness is improved and it can capture strong feature dependencies, even in the presence of imbalanced and label-scarce data. We validate our model’s effectiveness on three publicly available datasets i.e. BoT-IoT, ToN-IoT, and WUSTL-IIoT. Proposed framework improves inference time by a factor of 104 over State-of-The-Art (SOTA) methods. It also achieves a precision, recall and F1-score of 99%, 93% and 95% respectively which are comparable to existing SOTA methods.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110753"},"PeriodicalIF":4.9,"publicationDate":"2025-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145267250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Single image dehazing using multi-enhanced image stacking and pyramidal fusion","authors":"Nisa A.K. ,&nbsp;Vishnukumar S. ,&nbsp;Abin P. Mathew","doi":"10.1016/j.compeleceng.2025.110737","DOIUrl":"10.1016/j.compeleceng.2025.110737","url":null,"abstract":"<div><div>Imaging systems often generate images with low contrast, particularly in adverse weather conditions such as haze and fog. The reduced perceptibility in these images is affected by the dispersion and absorption of the light due to particles in the atmospheric aerosols. Due to the absence of depth information, image dehazing poses a challenge. To address this, fusion-based methods are employed as they can integrate complementary features from multiple enhancement techniques, allowing for a more comprehensive and balanced enhancement of both local details and global contrast. However, challenges such as uneven illumination, artifacts, and the preservation of fine details persist in many dehazing scenarios, motivating the need for more robust solutions. A novel fusion-based approach that uses multiple image enhancement techniques to improve the perceptual quality of the dehazed image is proposed here. Initially the Sharpening Smoothing Image Filter (SSIF) is applied on the input image for performing the sharpening operation. Then an intermediate image stack consisting of six enhanced images is developed using the sharpened image by applying Gamma Correction with four different gamma values, Contrast Limited Adaptive Histogram Equalization (CLAHE) and Multi-Scale Retinex with Color Restoration (MSRCR). This intermediate stack generation ensures that each enhancement technique contributes uniquely to the final result. A pyramidal fusion is applied to the images in the intermediate stack using contrast, saturation, and well-exposedness to produce the dehazed image. The fusion process effectively minimizes local haze variations, preserves color fidelity, and enhances image contrast under varying lighting conditions. The experimental analysis using five different datasets shows that the proposed method outperforms both qualitatively and quantitatively compared to existing methods.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110737"},"PeriodicalIF":4.9,"publicationDate":"2025-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145267312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Benchmarking NIST LWC candidates over GF22FDx achieving 3.5 Tb/J and 4.4 Gbps for IoT applications","authors":"Islam Elsadek ,&nbsp;Ahmed Zaky Ghonem ,&nbsp;Sherif Abouzeid ,&nbsp;John Ross Wallrabenstein ,&nbsp;Erik MacLean ,&nbsp;Doug Gardner ,&nbsp;Sohrab Aftabjahani ,&nbsp;Rosario Cammarota ,&nbsp;Eslam Yahya Tawfik","doi":"10.1016/j.compeleceng.2025.110758","DOIUrl":"10.1016/j.compeleceng.2025.110758","url":null,"abstract":"<div><div>Cryptography is a cornerstone of Internet-of-Things (IoT) security, protecting sensitive data and ensuring the integrity of communication channels. Standard cryptography, designed for server environments with ample resources algorithms e.g. Advanced Encryption Standard (AES), may not be suitable for the resource-constrained devices prevalent in IoT nodes, edge computing, and Unmanned Aerial Vehicles (UAVs). National Institute of Standards and Technology (NIST) initiated a standardization process for Lightweight Cryptography (LWC) algorithms to address this challenge. Ten candidate algorithms were selected for the final round in 2021, and Ascon was announced as the recommended LWC in 2023 and then published as a Special Publication (SP 800-232) instead of a Federal Information Processing Standard (FIPS) standard. As an SP, the use of Ascon in IoT nodes is recommended but not mandatory. This flexibility allows for the consideration of alternative algorithms based on specific application needs of area, performance and/or energy efficiency. Therefore, it is essential to evaluate and rank other LWC candidates in each metric. In addition, exploring Hardware (HW), Software (SW), and HW/SW co-design architectures is vital in IoT to meet diverse security and computational needs. This work provides a complete silicon-based benchmarking for the final candidates using three different architectures of HW, SW and HW/SW. Benchmarking is conducted fairly by unifying architectures, optimizations, and fabrication over the same chip. CMOS Global Foundries (GF22FDx) technology is used for fabrication. Results show that HW enhances the throughput by up to 6 orders of magnitude and energy efficiency by up to 5 orders of magnitude compared to SW. Xoodyak stands out as a top performer in terms of energy efficiency, while Sparkle excels in throughput, and TinyJambu is the smallest in terms of area.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110758"},"PeriodicalIF":4.9,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266108","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"In-the-wild facial emotion recognition using relation-aware geometric features and CapsNet","authors":"Nidhi,&nbsp;Bindu Verma","doi":"10.1016/j.compeleceng.2025.110685","DOIUrl":"10.1016/j.compeleceng.2025.110685","url":null,"abstract":"<div><div>Occlusions and pose variations are key challenges in Facial Emotion Recognition (FER), affecting recognition accuracy, especially in uncontrolled environments. This paper presents a robust FER method, FMR-CapsNet (includes Facemesh mediapipe, ResNet50, and Capsule Neural Network), designed to address these issues. The proposed model employs the FaceMesh model for geometric feature extraction, utilizing facial blendshape scores to capture expression-related features even in side-facing and occluded images. A Euclidean Distance metric constructs a relation-aware distance matrix to encode spatial relationships between blendshape scores. To further refine features, transfer learning is applied using a pretrained Residual Network (ResNet50), followed by a Capsule Neural Network (CapsNet) to capture directional and spatial information, improving feature differentiation. Extensive experiments on three in-the-wild datasets— Real-world Affective Faces Database (RAF-DB), AffectNet, and FERPlus demonstrate that FMR-CapsNet significantly enhances FER performance, achieving 97.01% accuracy on RAF-DB, 71.12% on AffectNet, and 91.82% on FERPlus, outperforming state-of-the-art (SOTA) methods in handling occlusions and pose variations.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110685"},"PeriodicalIF":4.9,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266110","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The blueprint of a new fact-checking system: A methodology to enrich RAG systems with new generated datasets","authors":"Salvador Lopez-Joya,&nbsp;Jose A. Diaz-Garcia,&nbsp;M. Dolores Ruiz,&nbsp;Maria J. Martin-Bautista","doi":"10.1016/j.compeleceng.2025.110746","DOIUrl":"10.1016/j.compeleceng.2025.110746","url":null,"abstract":"<div><div>In an era where digital misinformation spreads rapidly, Artificial Intelligence (AI) has become a crucial tool for fact-checking. However, the effectiveness of AI in this domain is often limited by the availability of high-quality and scalable datasets to train and guide algorithms. In this paper, we introduce VERIFAID (VERIfication FAISS-based framework for fake news Detection), a novel framework that improves fact-checking through a Retrieval-Augmented Generation (RAG) system based on automatically generated and dynamically growing datasets. Our approach improves evidence retrieval by building a scalable knowledge base, reducing the reliance on manually annotated data. The system consists of three key modules: two dedicated to dataset creation and one inference module that integrates advanced language models, such as LLaMA, within the RAG paradigm. To validate our methodology, we provide technical specifications for both the system and the dataset, together with comprehensive evaluations in zero-shot fact-checking scenarios. The results demonstrate the efficiency and adaptability of our approach and its potential to improve AI-driven fact verification at scale.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110746"},"PeriodicalIF":4.9,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266179","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Policy-driven contextual risk evaluation in OAuth 2.0 authentication frameworks for AI chatbot-based RPA systems","authors":"Soonhong Kwon,&nbsp;Wooyoung Son,&nbsp;Jong-Hyouk Lee","doi":"10.1016/j.compeleceng.2025.110759","DOIUrl":"10.1016/j.compeleceng.2025.110759","url":null,"abstract":"<div><div>With the advent of smartphones, we can access internet services regardless of location. This shift in environment has led to the demand for services that can be utilized anytime, anywhere. Consequently, Robotic Process Automation (RPA) technology, which automates simple and repetitive tasks in industrial settings, is gaining significant attention. There is a growing trend to combine this with Artificial Intelligence (AI) chatbot technology to achieve full automation and handle higher-level tasks. However, when performing high-level tasks based on an AI chatbot-based RPA system, situations arise where the AI chatbot relies on the user’s judgment. In such scenarios, the absence of an appropriate mechanism or technology to perform identity verification between the user and the AI chatbot exposes the system to security threats like personal information leakage and system takeover. Accordingly, this paper proposes an OAuth 2.0 integrated authentication framework utilizing a context-based risk assessment approach. This framework aims to reduce the likelihood of security threats and minimize the scale of damage caused by such threats. It achieves this by enabling access control based on the user’s context while requiring the user to provide minimal information when utilizing the AI chatbot-based RPA system. More specifically, the proposed framework employs a risk assessment based on the sigmoid function, which accounts for sensitivity variations across different contexts. This approach enables sensitive adjustments to access permissions in response to contextual changes, rather than applying a fixed risk assessment. This demonstrates the framework’s capability to provide a trustworthy automated work environment through appropriate access control. Specifically, the proposed risk assessment formula quantitatively analyzes sensitivity changes for each contextual variable through mathematical interpretation. Based on this, it structurally derives the correlation between risk scores and policies. Furthermore, experimental results confirm consistency between policy flow and risk assessment, such as issuing ‘Full Access Tokens’ in normal situations and applying Access Denied in high-risk situations. Furthermore, using data flow diagrams and STRIDE, potential security threats within the proposed framework were modeled. Simulation of actual security threats demonstrated the framework’s ability to mitigate these threats, with an average latency of 9.22ms and memory usage of 64.00MB required for threat response. This empirically demonstrates that the proposed framework is a valid authentication structure capable of simultaneously achieving real-time performance, security, and lightweight characteristics even in AI-based automated environments.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110759"},"PeriodicalIF":4.9,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266185","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ETRC-net: Efficient transformer for grading renal cell carcinoma in histopathological images","authors":"Mohsin Raza ,&nbsp;Umme E Farwa ,&nbsp;Md Ariful Islam Mozumder ,&nbsp;Joo Mon-il ,&nbsp;Hee-Cheol Kim","doi":"10.1016/j.compeleceng.2025.110747","DOIUrl":"10.1016/j.compeleceng.2025.110747","url":null,"abstract":"<div><div>Renal cell carcinoma (RCC), the most prevalent form of kidney cancer, accounts for nearly 85 % of kidney cancer-related deaths. Manual diagnosis of RCC from histopathology images relies heavily on the expertise of pathologists, often leading to variability in results. Although deep learning methods have been explored for disease diagnosis, research on RCC remains limited, and existing approaches are insufficient for accurate grading. Since each RCC stage requires a distinct treatment plan, reliable grading is crucial, as errors can result in inappropriate therapies and poor patient outcomes. To address this challenge, we propose the Efficient Transformer for Renal Classification Network (ETRC<img>Net), a novel deep learning framework specifically designed for accurate RCC classification from histopathology images. ETRC<img>Net combines EfficientNet with Squeeze-and-Excitation (SE) blocks for enhanced feature representation and a customized Vision Transformer encoder to capture global context and long-range dependencies. The SE blocks adaptively recalibrate channel-wise responses, enabling the model to focus on relevant features while suppressing less informative ones. We evaluate ETRC<img>Net on the Kasturba Medical College (KMC) dataset, achieving 94.37 % accuracy, 94.54 % precision, 94.37 % recall, and an F1-score of 94.37 %. On the Lung and Colon dataset, it further demonstrates superior generalization with 99.92 % accuracy, 99.64 % precision, 99.71 % recall, and a 99.80 % F1-score. Compared to state-of-the-art methods, ETRC<img>Net delivers higher accuracy with fewer trainable parameters and lower computational cost. Its efficiency and scalability make itfor resource constrained clinical environments, offering a robust and intelligent solution for early RCC diagnosis.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110747"},"PeriodicalIF":4.9,"publicationDate":"2025-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266180","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}