Array最新文献

{"title":"Accurate industrial anomaly detection with efficient multimodal fusion","authors":"Dinh-Cuong Hoang ,&nbsp;Phan Xuan Tan ,&nbsp;Anh-Nhat Nguyen ,&nbsp;Ta Huu Anh Duong ,&nbsp;Tuan-Minh Huynh ,&nbsp;Duc-Manh Nguyen ,&nbsp;Minh-Duc Cao ,&nbsp;Duc-Huy Ngo ,&nbsp;Thu-Uyen Nguyen ,&nbsp;Khanh-Toan Phan ,&nbsp;Minh-Quang Do ,&nbsp;Xuan-Tung Dinh ,&nbsp;Van-Hiep Duong ,&nbsp;Ngoc-Anh Hoang ,&nbsp;Van-Thiep Nguyen","doi":"10.1016/j.array.2025.100512","DOIUrl":"10.1016/j.array.2025.100512","url":null,"abstract":"<div><div>Industrial anomaly detection is critical for ensuring quality and efficiency in modern manufacturing. However, existing deep learning models that rely solely on red-green-blue (RGB) images often fail to detect subtle structural defects, while most RGB-depth (RGBD) methods are computationally heavy and fragile in the presence of missing or noisy depth data. In this work, we propose a lightweight and real-time RGBD anomaly detection framework that not only refines per-modality features but also performs robust hierarchical fusion and tolerates missing inputs. Our approach employs a shared ResNet-50 backbone with a Modality-Specific Feature Enhancement (MSFE) module to amplify texture and geometric cues, followed by a Hierarchical Multi-Modal Fusion (HMM) encoder for cross-scale integration. We further introduce a curriculum-based anomalous feature generator to produce context-aware perturbations, training a compact two-layer discriminator to yield precise pixel-level normality scores. Extensive experiments on the MVTec Anomaly Detection (MVTec-AD) dataset, the Visual Anomaly (VisA) dataset, and a newly collected RealSense D435i RGBD dataset demonstrate up to 99.0% Pixel-level Area Under the Receiver Operating Characteristic Curve (P-AUROC), 99.6% Image-level AUROC (I-AUROC), 82.6% Area Under the Per-Region Overlap (AUPRO), and 45 frames per second (FPS) inference speed. These results validate the effectiveness and deployability of our approach in high-throughput industrial inspection scenarios.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100512"},"PeriodicalIF":4.5,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145118952","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Drone detection in airport environments: A literature review","authors":"Sanderson Oliveira de Macedo ,&nbsp;Mauro Caetano ,&nbsp;Ronaldo Martins da Costa","doi":"10.1016/j.array.2025.100511","DOIUrl":"10.1016/j.array.2025.100511","url":null,"abstract":"<div><div>The increasing use of drones in airport airspace presents a serious challenge to safety and efficiency. Incidents involving unmanned aerial vehicles can cause delays, flight cancellations, and collision risks, raising concerns among airport officials, travelers, and other aviation stakeholders. This study aims to systematically analyze the main drone detection techniques used in airports, identifying research gaps, advantages, and limitations of each method while also highlighting future directions to improve airspace security. Kitchenham’s systematic review method was used, with searches carried out from 2014 to 2025. After screening titles and abstracts and applying inclusion criteria, 25 publications were thoroughly assessed. The analysis shows that while radar systems provide the longest detection range (<span><math><mrow><mo>&gt;</mo><mn>10</mn></mrow></math></span> km) and radio frequency methods achieve the highest classification accuracy (<span><math><mo>∼</mo></math></span>99%), they often come with higher costs. In comparison, camera-based systems can reach high precision (<span><math><mo>&gt;</mo></math></span>90%) at speeds up to 170 FPS, and multimodal solutions show the greatest potential for robustness, with positioning errors below 1.5% of the detection range. Although technical and operational challenges still exist, the combined use of various methods and machine learning techniques shows promise for improving the accuracy and reliability of drone detection at airports.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100511"},"PeriodicalIF":4.5,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145217797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Privacy protection in steel supply chain via blockchain and attribute-based proxy re-encryption","authors":"Jie Gao ,&nbsp;Xiaohong Zhang ,&nbsp;Shaojiang Zhong","doi":"10.1016/j.array.2025.100514","DOIUrl":"10.1016/j.array.2025.100514","url":null,"abstract":"<div><div>With the rapid advancement of globalization and digitalization, the steel supply chain has become increasingly complex, raising significant concerns regarding data privacy and security. Traditional supply chain management models face critical limitations in ensuring data confidentiality, transparency, and efficient data sharing, rendering them inadequate for the evolving demands of the modern steel industry. To address these challenges, this paper proposes a blockchain-based mechanism for secure and private data sharing within the steel supply chain. The proposed framework integrates attribute-based proxy re-encryption (ABPRE) with a blockchain-enabled storage and sharing scheme, which not only ensures user data privacy but also enhances system execution efficiency. By leveraging ABPRE, data owners can share encrypted information directly with other authorized users without the need for repeated encryption, uploading, or downloading, significantly reducing access pressure on cloud storage. Compared to conventional cloud-based solutions, our scheme demonstrates superior security capabilities and practical efficiency. The integration of InterPlanetary File System (IPFS) further optimizes decentralized data storage and access, while the database-level privacy protection mechanism maintains data integrity during storage and transmission. Additionally, the use of smart contracts automates encryption and decryption operations, improving flexibility in access control management. Experimental results show that the proposed scheme effectively preserves data privacy while outperforming existing ABPRE schemes in terms of both security and computational efficiency. Its application to steel supply chain management demonstrates strong practical viability and technical robustness.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100514"},"PeriodicalIF":4.5,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145097765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Transforming ERP systems with collaborative AI: Paving the path to strategic growth and sustainability","authors":"Muhammad Adeel Mannan ,&nbsp;Abdul Khalique Shaikh ,&nbsp;Abdullah Ayub Khan ,&nbsp;Kaamran Raahemifar ,&nbsp;Mohamad Afendee Mohamed","doi":"10.1016/j.array.2025.100517","DOIUrl":"10.1016/j.array.2025.100517","url":null,"abstract":"<div><div>Collaborative AI is one of the essential part of any system now a days for sustainability of any system especially ERP. Tis research study investigates the collaborative AI integration to ERP systems implementation and its impact on it while business growth and long term organizational sustainability. For any businesses encounter critical issues with digitalization, compliance and optimization in today's innovative and intelligent society. Recent advancements in the suggested collaborative AI-ERP frameworks are investigate in this article, predictive analytics for long-term sustainability in decision-making, natural language based interfaces for user engagement, and machine learning implementation for process improvement. We analyze key areas of implementation problems and potential future paths that could impact both scholarly facts and commercial uses. By strategically formalize these resources, and enabling the circular economy, AI-enhanced system ERP implementation may support growth and sustainability goals and increase te operational efficiency.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100517"},"PeriodicalIF":4.5,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145097764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Adaptive multi-scale spatio-temporal convolutional network with reinforcement learning for dynamic lane-level traffic flow prediction","authors":"Xiaohui Yang ,&nbsp;Shaowei Sun ,&nbsp;Mingzhou Liu","doi":"10.1016/j.array.2025.100513","DOIUrl":"10.1016/j.array.2025.100513","url":null,"abstract":"<div><div>This paper presents an Adaptive Multi-Scale Spatio-Temporal Convolutional Network and Reinforcement Learning Collaborative Optimization Lane-Level Traffic Flow Prediction Model (AST-RLM), designed to address the challenges posed by the sudden changes in microscopic driving behaviors and spatio-temporal dependencies in complex urban environments. The model achieves high-precision lane-level traffic flow prediction through dynamic graph construction mechanisms, heterogeneous perception-based multi-scale convolutional networks, and a DQN-based collaborative optimization framework. Experimental results demonstrate that AST-RLM performs exceptionally well on real-world datasets from multiple cities, containing over 10,000 lanes. The average absolute error (MAE) during the evening peak is as low as 0.033, a 38.9 % reduction compared to GraphWaveNet. The root mean square error (RMSE) for 30-min predictions is 3.98, outperforming existing models like ST-MetaNet, and the model maintains 92.4 % stability even in extreme weather conditions. Notably, during sudden events like traffic accidents, the dynamic graph module adapts in real-time to changes in topology, reducing prediction errors by 26.7 %–30.9 %, significantly improving the model's robustness and responsiveness in complex dynamic scenarios. Furthermore, AST-RLM's multi-agent reinforcement learning deployment on edge devices achieves a convergence speed 3.6 times faster than GC-RL, validating its efficiency and feasibility in real-world traffic systems.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100513"},"PeriodicalIF":4.5,"publicationDate":"2025-09-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145097792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating clustering with evolutionary feature selection using ENORA and SToWVector","authors":"Alexander José Mackenzie-Rivero ,&nbsp;Rodrigo Martínez-Béjar ,&nbsp;Hilarión José Vegas-Meléndez","doi":"10.1016/j.array.2025.100508","DOIUrl":"10.1016/j.array.2025.100508","url":null,"abstract":"<div><div>The rapid growth of textual data from sources such as social media, blogs, and digital libraries has intensified the demand for scalable and semantically informed classification methods. This study introduces a hybrid framework that integrates unsupervised clustering, evolutionary feature selection, and semantic interpretation to enhance automatic text classification. The approach combines the SToWVector representation with a Multi-Objective Evolutionary Search (MOES) strategy optimized through the ENORA algorithm, while employing the NaiveBayesMultinomial classifier for evaluation. Semantic interpretation is incorporated via ontological reasoning, enabling the model to capture latent conceptual relationships among terms and thereby complement both the clustering and feature selection processes. Experimental evaluations on benchmark and large-scale datasets (SMS Spam and Euronews) demonstrate the robustness of the framework, including a scenario in which 100% accuracy was achieved. The proposed method outperforms traditional models and achieves competitive results against deep learning-based classifiers. These findings underscore the framework’s adaptability and effectiveness in managing high-dimensional unstructured text, while preserving interpretability through symbolic reasoning.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100508"},"PeriodicalIF":4.5,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145048534","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring the role of generative AI in enhancing cybersecurity in software development life cycle","authors":"Hussein A. Al-Hashimi ,&nbsp;Rafiq Ahmad Khan ,&nbsp;Hathal S. Alwageed ,&nbsp;Asaad M. Algarni ,&nbsp;Sarra Ayouni ,&nbsp;Alaa Omran Almagrabi","doi":"10.1016/j.array.2025.100509","DOIUrl":"10.1016/j.array.2025.100509","url":null,"abstract":"<div><h3>Context</h3><div>The rapid integration of Generative AI (GenAI) technologies in various sectors has introduced new opportunities and challenges. One of the areas where GenAI is gaining prominence is cybersecurity, particularly within the Software Development Life Cycle (SDLC). As cyber threats evolve, there is a growing need to explore innovative solutions to mitigate vulnerabilities during software development.</div></div><div><h3>Objectives</h3><div>This study investigates the role of GenAI in enhancing cybersecurity in the SDLC. It examines current security practices, recent advancements in AI-driven security solutions, and the potential of GenAI to strengthen threat detection, vulnerability management, and risk mitigation. Additionally, the research identifies key opportunities and challenges associated with integrating GenAI into SDLC processes, highlighting its implications for secure software development and future industry practices.</div></div><div><h3>Methods</h3><div>This research employs a mixed-methods approach to investigate the role of GenAI in cybersecurity. Specifically, it combines a Systematic Literature Review (SLR) with questionnaire-based data collection targeting software development and cyber defense experts. The SLR aims to identify prevailing themes and gaps, while the questionnaire gathers insights from IT professionals about their experiences and perspectives on GenAI systems.</div></div><div><h3>Results</h3><div>Our research shows that GenAI technology enhances SDLC security by supporting development through vulnerability detection, threat modeling, secure coding practices, and incident response. However, our review shows that AI adoption introduces ethical risks alongside reliability issues with AI-created results and challenges to integrate it into standard development methods.</div></div><div><h3>Conclusion</h3><div>The integration of GenAI into the SDLC offers significant potential for enhancing cybersecurity. While challenges such as algorithm transparency and the need for skilled professionals remain, the benefits of AI in proactive threat detection and response make it a promising tool for future cybersecurity strategies in software development.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100509"},"PeriodicalIF":4.5,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145097763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A retrieval augmented generation based optimization approach for medical knowledge understanding and reasoning in large language models","authors":"Yingshuai Wang ,&nbsp;Yanli Wan ,&nbsp;Xingyun Lei ,&nbsp;Qingkun Chen ,&nbsp;Hongpu Hu","doi":"10.1016/j.array.2025.100504","DOIUrl":"10.1016/j.array.2025.100504","url":null,"abstract":"<div><div>Based on the existing Retrieval Augmented Generation (RAG) technology, this study proposes innovative solution to better address the hallucination issues of current large language models. By optimizing data processing, prompt engineering, and multi-retriever fusion, it resolves issues such as semantic capture bias, inaccurate context retrieval, information redundancy, hallucination generation, and length limitations. Data processing focuses on text cleaning, disambiguation, and removing redundant information to enhance consistency. Prompt engineering aids the model in better understanding the task. The adaptive weight fusion of sparse and dense retrievers improves context retrieval accuracy. Experiments conducted on the CCKS-TCMBench dataset for medical knowledge understanding and semantic reasoning show that the optimized model significantly outperforms the baseline across all evaluation metrics.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100504"},"PeriodicalIF":4.5,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145048533","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Clustering microbial material on agar plates: A modular system approach","authors":"Michal Cicatka ,&nbsp;Radim Burget ,&nbsp;Jan Karasek ,&nbsp;Jan Lancos","doi":"10.1016/j.array.2025.100503","DOIUrl":"10.1016/j.array.2025.100503","url":null,"abstract":"<div><div>Analyzing microbial colonies on agar plates is a critical task in microbiology, with applications spanning research, diagnostics, and industry. Despite advances in automated systems, challenges remain in accurately segmenting and clustering colonies due to variability in their appearance and distribution. To address these challenges, we present a fully automated modular system for colony segmentation and clustering, combining state-of-the-art deep learning for segmentation and machine learning for cluster count prediction and clustering.</div><div>The main contribution of this work is the proposition, development and evaluation of a novel system, which achieved a V-measure of 0.532 under real-world conditions, improving to 0.727 with ideal segmentation and cluster counts, setting a new benchmark for microbiological analysis. At the core of the system we propose AgarNet, an Attention U-Net-based architecture combined with an EfficientNetB4 backbone, which achieved an F1-score of 0.906 for segmentation. We also introduce a new BRUKERCLUSTER dataset, which is one of the largest and most diverse annotated resources of its kind, featuring expert-annotated images from a controlled cultivation. By combining the dataset, robust segmentation, accurate cluster count prediction, and effective clustering, the proposed system delivers a scalable solution for advancing automated microbial colony analysis.</div><div>These results establish a new benchmark for automated colony clustering in microbiology, and by releasing the unique BRUKERCLUSTER dataset, we provide a valuable tool for future advancements in automated colony analysis.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"28 ","pages":"Article 100503"},"PeriodicalIF":4.5,"publicationDate":"2025-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145217806","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hybrid framework for Remaining Useful Life (RUL) prediction of rolling bearing faults","authors":"Ali Saeed ,&nbsp;Muazzam A. Khan ,&nbsp;Salman Jan ,&nbsp;Arslan Shaukat ,&nbsp;M. Usman Akram ,&nbsp;Toqeer Ali Syed ,&nbsp;Adeel M. Syed","doi":"10.1016/j.array.2025.100498","DOIUrl":"10.1016/j.array.2025.100498","url":null,"abstract":"<div><div>Remaining Useful Life (RUL) prediction is critical for preventing catastrophic failures in industrial systems, enabling efficient maintenance scheduling and resource optimization. This paper presents a novel hybrid framework for RUL prediction that integrates advanced feature extraction with state-of-the-art deep learning methods. The proposed framework employs Modified Multiscale Permutation Entropy (MMPE) to compute a robust Health Indicator (HI) representing the system’s degradation behavior. A regression transformer model, trained on the FEMTO dataset and validated on the CWRU dataset, leverages the HI to capture complex temporal dependencies and predict RUL with high accuracy. Experimental results demonstrate the superior performance of the proposed approach, achieving a Mean Squared Error (MSE) of <span><math><mrow><mn>3</mn><mo>.</mo><mn>6</mn><mo>×</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</mo><mn>5</mn></mrow></msup></mrow></math></span> and Mean Absolute Error (MAE) of <span><math><mrow><mn>5</mn><mo>.</mo><mn>48</mn><mo>×</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</mo><mn>3</mn></mrow></msup></mrow></math></span>, significantly outperforming existing methods. The framework’s ability to generalize across datasets and operating conditions highlights its applicability for real-world industrial settings, offering a robust and interpretable solution for predictive maintenance.</div></div>","PeriodicalId":8417,"journal":{"name":"Array","volume":"27 ","pages":"Article 100498"},"PeriodicalIF":4.5,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144931613","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}