Publications

@article{barbu_framework_2025,

title = {Framework Design for Reinforcing the Potential of XR Technologies in Transforming Inclusive Education},

author = {M. Barbu and D. -D. Iordache and I. Petre and D. -C. Barbu and L. Bajenaru},

url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85217742383&doi=10.3390%2Fapp15031484&partnerID=40&md5=9ff9c99c76855723172055c73049fb5a},

doi = {10.3390/app15031484},

issn = {20763417 (ISSN)},

year  = {2025},

date = {2025-01-01},

journal = {Applied Sciences (Switzerland)},

volume = {15},

number = {3},

abstract = {This study presents a novel approach to inclusive education by integrating augmented reality (XR) and generative artificial intelligence (AI) technologies into an immersive and adaptive learning platform designed for students with special educational needs. Building upon existing solutions, the approach uniquely combines XR and generative AI to facilitate personalized, accessible, and interactive learning experiences tailored to individual requirements. The framework incorporates an intuitive Unity XR-based interface alongside a generative AI module to enable near real-time customization of content and interactions. Additionally, the study examines related generative AI initiatives that promote inclusion through enhanced communication tools, educational support, and customizable assistive technologies. The motivation for this study arises from the pressing need to address the limitations of traditional educational methods, which often fail to meet the diverse needs of learners with special educational requirements. The integration of XR and generative AI offers transformative potential by creating adaptive, immersive, and inclusive learning environments. This approach ensures real-time adaptability to individual progress and accessibility, addressing critical barriers such as static content and lack of inclusivity in existing systems. The research outlines a pathway toward more inclusive and equitable education, significantly enhancing opportunities for learners with diverse needs and contributing to broader social integration and equity in education. © 2025 Elsevier B.V., All rights reserved.},

note = {Publisher: Multidisciplinary Digital Publishing Institute (MDPI)},

keywords = {Adaptive Learning, Adversarial machine learning, Artificial intelligence technologies, Augmented Reality, Contrastive Learning, Educational Technology, Extended reality (XR), Federated learning, Framework designs, Generative adversarial networks, Immersive, immersive experience, Immersive learning, Inclusive education, Learning platform, Special education needs},

pubstate = {published},

tppubtype = {article}

}

@article{wang_mobile-driven_2025,

title = {Mobile-Driven Deep Learning Algorithm for Personalized Clothing Design Using Multi-Feature Attributes},

author = {Z. Wang and A. Aris and P. Zhang},

url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-105017860148&doi=10.3991%2Fijim.v19i18.57239&partnerID=40&md5=de3ca359dd178d8ea59cf8da73a9c486},

doi = {10.3991/ijim.v19i18.57239},

issn = {18657923 (ISSN)},

year  = {2025},

date = {2025-01-01},

journal = {International Journal of Interactive Mobile Technologies},

volume = {19},

number = {18},

pages = {146–160},

abstract = {Personalized fashion recommendation systems face significant challenges in balancing accurate style prediction, real-time mobile performance, and user privacy compliance. This study presents StyleFitNet, a novel mobile-driven deep learning framework that integrates multiple user feature attributes, including body measurements, fabric preferences, and temporal style evolution, to generate personalized clothing designs. The hybrid convolutional neural networks (CNNs)-recurrent neural networks (RNNs) architecture addresses key limitations of conventional recommendation systems by simultaneously processing spatial features and sequential preference patterns. A comprehensive evaluation demonstrates the system’s superiority in recommendation accuracy, design diversity, and user satisfaction compared to existing approaches. The implementation features GDPR-compliant data handling and a 3D virtual fitting room, significantly reducing return rates while maintaining robust privacy protections. Findings highlight the model’s ability to adapt to evolving fashion trends while preserving individual style preferences, offering both technical and business advantages for e-commerce platforms. The study concludes that StyleFitNet establishes a new standard for artificial intelligence (AI)-driven fashion recommendations, successfully merging advanced personalization with ethical data practices. Key implications include the demonstrated viability of hybrid deep learning models for mobile deployment and the importance of temporal analysis in preference modelling. Future research directions include cross-cultural validation and the integration of generative AI for enhanced visualization. © 2025 Elsevier B.V., All rights reserved.},

note = {Publisher: International Federation of Engineering Education Societies (IFEES)},

keywords = {Clothing design, Convolutional Neural Networks, Data privacy, Data visualization, Deep learning, E-Learning, Electronic commerce, Fashion design, Feature attributes, Hosiery manufacture, Learning algorithms, Learning platform, Learning systems, Mobile Learning, Mobile learning platform, Mobile-driven deep learning, Multi-feature attribute, multi-feature attributes, Multifeatures, Personalized clothing design, Personalized clothings, StyleFitNet, Textiles, Virtual Reality},

pubstate = {published},

tppubtype = {article}

}