International Journal of Internet of Things and Smart Computing Environment

A Blockchain-driven Framework for Secure Big Data Analytics through Data Science Integration

Ahamad Shariful Alam, Bidhita Islam Chowdhury — Mon, 01 Jun 2026 00:00:00 +0000

The rapid growth of big data technologies has transformed modern industries by enabling large-scale data processing, predictive analytics, and intelligent decision-making. However, centralized big data architectures face major security and privacy challenges, including unauthorized access, data tampering, and a lack of transparency. Blockchain technology offers decentralized, immutable, and transparent mechanisms that can enhance the security and integrity of big data analytics systems. This research article investigates the integration of blockchain and data science techniques for secure big data analytics. A conceptual framework combining distributed ledger technology, machine learning algorithms, and cloud-based big data infrastructure is proposed. The study evaluates the performance of the integrated model in terms of security, scalability, transparency, and analytical efficiency. Experimental findings demonstrate that blockchain integration significantly improves data integrity, authentication, and trust management while maintaining acceptable computational overhead. The study also discusses implementation challenges, future research directions, and real-world applications in healthcare, finance, smart cities, and IoT ecosystems.

Comparative Analysis of Classical and Deep Learning Methods for Robust Face Detection Systems

Al Ani Mohammed Nsaif Mustafa — Wed, 13 May 2026 00:00:00 +0000

Face detection is a fundamental task in computer vision and artificial intelligence, playing a critical role in applications such as surveillance, biometric authentication, and human-computer interaction. This study presents a comparative analysis of classical and deep learning-based face detection methods to evaluate their effectiveness under real-world conditions. The classical Viola-Jones algorithm is implemented as a baseline due to its computational efficiency and real-time performance capabilities. In contrast, a deep learning-based approach using the multi-task cascaded convolutional neural network (MTCNN) is employed to demonstrate the advantages of data-driven feature learning. The performance of both methods is evaluated using standard benchmark datasets, including WIDER FACE and FDDB, under challenging conditions such as occlusion, illumination variation, and scale diversity. Experimental results indicate that while the Viola-Jones method offers faster processing time, it suffers from reduced accuracy in complex environments. Conversely, the MTCNN model achieves higher detection accuracy, precision, and robustness, albeit at a higher computational cost. The findings highlight the trade-off between efficiency and accuracy and emphasize the importance of selecting appropriate detection methods based on application requirements. This study contributes to the development of more reliable and efficient face detection systems for real-world deployment.

Zero Trust in the Age of AI: An Adaptive Cybersecurity Architecture for Resilient Distributed Systems

Davidson B, Ejekwu Obunezi — Thu, 14 May 2026 00:00:00 +0000

The rapid expansion of distributed computing, cloud-native architectures, Internet of Things (IoT), and artificial intelligence (AI)-driven applications has significantly expanded the cyber-attack surface of modern organizations. Traditional perimeter-based security models are increasingly ineffective against sophisticated adversaries leveraging automation, polymorphic malware, and social engineering at scale. High-profile incidents such as the SolarWinds Orion supply chain compromise and ransomware campaigns like WannaCry demonstrate systemic weaknesses in implicit trust assumptions within enterprise networks. In response, Zero Trust Architecture (ZTA) has emerged as a transformative cybersecurity paradigm, emphasizing continuous verification, least-privilege access, and context-aware authentication. This article proposes an adaptive zero trust framework augmented with AI-driven behavioral analytics and dynamic risk scoring to enhance resilience in distributed systems. An integrated architecture was formulated combining identity-centric access control, micro-segmentation, encrypted telemetry pipelines, and machine learning-based anomaly detection. A hybrid evaluation methodology incorporating simulated attack scenarios and performance benchmarking is presented. Findings indicate that adaptive Zero Trust reduces lateral movement risk by over 60% in controlled environments while maintaining acceptable system latency overhead. The study contributes a scalable design blueprint for organizations transitioning from legacy perimeter defenses to intelligent, self-adjusting cybersecurity ecosystems. By integrating Zero Trust principles with AI-enabled situational awareness, the proposed framework strengthens confidentiality, integrity, and availability in contemporary digital infrastructures.

Impact of 5G Connectivity on Industrial Automation and Smart Factories

Mission Franklin — Mon, 02 Feb 2026 00:00:00 +0000

The rapid evolution of industrial automation and the emergence of smart factories have placed unprecedented demands on communication networks, requiring ultra-low latency, high reliability, and massive device connectivity. This paper investigates the role of 5G connectivity as a catalyst for the next generation of industrial automation, particularly within the context of Industry 5.0, where human-centric, resilient, and highly connected manufacturing systems are emphasized. By combining a comprehensive literature review, detailed case analyses of two smart factory pilots, and semi-structured interviews with industry experts and network engineers, the study examines the core enabling features of 5G, including ultra-reliable low-latency communications (URLLC), enhanced mobile broadband (eMBB), massive machine-type communications (mMTC), and network slicing. The findings demonstrate that 5G significantly enhances industrial operations by reducing latency from approximately 20 milliseconds (Wi-Fi) to around 1 millisecond, supporting high-bandwidth applications exceeding 200 Mbps, and providing reliable, deterministic communication across densely connected devices. These improvements facilitate real-time robotic control, collaborative human-machine interaction, augmented reality-assisted training, predictive maintenance, and high-resolution process monitoring. The study also identifies persistent challenges, including integration with legacy industrial control systems (ICS) and operational technology (OT), network slicing management, cybersecurity vulnerabilities, and regulatory compliance. Based on these insights, the paper provides a roadmap for 5G adoption in industrial settings, emphasizing best practices for deployment, integration, and operational optimization. Furthermore, it highlights key avenues for future research, including the development of secure and adaptive network management frameworks, advanced integration of edge computing, and strategies to ensure interoperability with existing industrial infrastructures. Overall, this study highlights the transformative potential of 5G in enabling highly responsive, flexible, and resilient smart factories, while providing a structured framework to guide industrial stakeholders in effectively leveraging 5G technologies.

Design and Implementation of an AI-powered CT Scan Report Analyzer for Web and Mobile Platforms

Archana Kale, Om Kharche, Atharva Ombase, Vedant Thakare, Anunay Patil — Sat, 07 Mar 2026 00:00:00 +0000

Brain tumors are one of the most life-threatening neurological disorders, requiring early and accurate diagnosis for effective treatment planning. Computed tomography (CT) remains one of the most reliable and widely accessible imaging modalities for brain screening, especially in emergency and resource-limited healthcare settings. However, manual interpretation of CT images is highly subjective and prone to variability among radiologists, leading to delays in diagnosis and potential misclassification of tumor regions. This paper presents a highly scalable, intelligent, and explainable AI-based CT scan analyzer (Web/App) that leverages hybrid deep learning and modern web technologies for automated tumor detection and visualization. The proposed system combines three major deep learning paradigms—Convolutional neural networks (CNN) for spatial feature extraction, genetic algorithm (GA) for feature selection and optimization, and bidirectional long short-term memory (BiLSTM) for temporal and contextual classification. The model is deployed via a ReactJS frontend and a Flask-TensorFlow backend, creating a responsive web application capable of real-time inference, visualization through Grad-CAM heatmaps, and secure data handling via MongoDB. This study consolidates literature between 2022 and 2025 and discusses the methodology, architecture, explainability, deployment, and clinical implications of integrating hybrid AI systems in radiological diagnostics.