{"title":"Optimizing medical visual question answering: Evaluating the impact of enhanced images, augmented training data, and model selection","authors":"Ali Jaber Almalki","doi":"10.1002/itl2.588","DOIUrl":"https://doi.org/10.1002/itl2.588","url":null,"abstract":"<p>Visual question answering (VQA) has an interesting application in clinical decision support and enables clinicians to extract information from medical images through natural language queries. However, the limited nature of the datasets makes it particularly difficult to develop effective VQA models for the medical profession. The aim of this study was to overcome these obstacles by formally testing methods for data enhancement and model optimization. Specifically, we merged two medical VQA datasets, applied image preprocessing techniques, examined several state-of-the-art model architectures, and extensively trained the best-performing model on the enhanced data. The results showed that training the VGG16-LSTM model on sharper images than the merged dataset resulted in a significant performance improvement due to extending the training time to 200, with F1 scores of the training set 0.9674.</p>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"8 2","pages":""},"PeriodicalIF":0.9,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143496934","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":"Security Enhancement in 5G Networks by Identifying Attacks Using Optimized Cosine Convolutional Neural Network","authors":"Premalatha Santhanamari, Vijayakumar Kathirgamam, Lakshmisridevi Subramanian, Thamaraikannan Panneerselvam, Rathish Chirakkal Radhakrishnan","doi":"10.1002/itl2.70003","DOIUrl":"https://doi.org/10.1002/itl2.70003","url":null,"abstract":"<div>\u0000 \u0000 <p>The exponential growth of 5G networks has introduced advanced capabilities but also heightened susceptibility to sophisticated cyberattacks. To address this, a robust and optimized security framework is proposed, leveraging a Cosine Convolutional Neural Network (CCNN) for attack detection. By emphasizing angular correlations in data, the CCNN improves feature extraction by substituting cosine similarity-based adjustments for conventional convolution processes. To maximize the CCNN's performance, the Exponential Distribution Optimizer (EDO) is employed optimize CCNN. The optimal configuration of CCNN is achieved using EDO's probabilistic search mechanism, which is inspired by exponential distribution and helps to maintain a balanced exploration-exploitation strategy. This integrated approach significantly improves detection accuracy, robustness, and scalability while maintaining low computational overhead. Comprehensive evaluations demonstrate the model's efficacy in identifying diverse attack patterns in 5G networks, outperforming conventional methods. The proposed framework establishes a new benchmark for secure, intelligent 5G infrastructures, contributing to the advancement of cybersecurity in next-generation networks. The introduced approach attains higher accuracy of 99%.</p>\u0000 </div>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"8 2","pages":""},"PeriodicalIF":0.9,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143496937","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":"Trust-Based Reliability Scheme for Secure Data Sharing With Internet of Vehicles Networks","authors":"Arpit Jain, Ashok Kumar, Mahadev, Jitendra Kumar Chaudhary, Saurabh Singh","doi":"10.1002/itl2.70000","DOIUrl":"https://doi.org/10.1002/itl2.70000","url":null,"abstract":"<div>\u0000 \u0000 <p>With the Internet of Things (IoT) increasingly integrated into vehicles, drivers and passengers can access information anywhere, anytime. As the number of connected vehicles increases, new requirements for vehicular networks arise, including securing, robust, and scalable communication between vehicles and pedestrians. It encompasses the communication between vehicles and infrastructure, as well as the communication between vehicles and pedestrians. A real-time exchange of road condition information can be achieved through this method. An IoV network rogue node detection scheme is presented here using a combined trust model. Using a trust-based security algorithm, vehicles are assessed for trustworthiness and rogue vehicles are identified. Communication and data reliability are used to calculate direct Trust while neighboring vehicles' cooperation is used to calculate recommendation trust. According to simulation results, the proposed model is highly accurate and maintains a low detection delay while protecting evaluation data privacy. Trust updates are more efficient using this model, and malicious vehicles are detected more quickly than traditional schemes.</p>\u0000 </div>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"8 2","pages":""},"PeriodicalIF":0.9,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143496935","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":"Edge Computing Enables Assessment of Student Community Building: An Emotion Recognition Method Based on TinyML","authors":"Shuo Liu","doi":"10.1002/itl2.645","DOIUrl":"https://doi.org/10.1002/itl2.645","url":null,"abstract":"<div>\u0000 \u0000 <p>Deep network-based video sentiment analysis is crucial for online evaluation tasks. However, these deep models are difficult to run on intelligent edge devices with limited computing resources. In addition, video data are susceptible to lighting interference, distortion, and background noise, which severely limits the performance of facial expression recognition. To relieve these issues, we develop an effective multi-scale semantic fusion tiny machine learning (TinyML) model based on a spatiotemporal graph convolutional network (ST-GCN) which enables robust expression recognition from facial landmark sequences. Specifically, we construct regional-connected graph data based on facial landmarks which are collected from cameras on different mobile devices. In existing spatiotemporal graph convolutional networks, we leverage the multi-scale semantic fusion mechanism to mine the hierarchical structure of facial landmarks. The experimental results on CK+ and online student community assessment sentiment analysis (OSCASA) dataset confirm that our approach yields comparable results.</p>\u0000 </div>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"8 2","pages":""},"PeriodicalIF":0.9,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143481531","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}
Dr. B. D. Parameshachari, Dr. Danilo Pelusi, Dr. Bhargavi Goswami
{"title":"AI-Driven Big Data Analytics for Mobile Healthcare","authors":"Dr. B. D. Parameshachari, Dr. Danilo Pelusi, Dr. Bhargavi Goswami","doi":"10.1002/itl2.624","DOIUrl":"https://doi.org/10.1002/itl2.624","url":null,"abstract":"<div>\u0000 \u0000 <p>This special issue highlights the key insights into the emerging trends and challenges of artificial intelligence (AI) driven big data analytics in mobile healthcare. The selected twenty-six articles, which include two letters, explore innovative applications and advancements in AI-driven healthcare technologies. The contributions of selected articles are divided into four main themes: Medical technology in healthcare communication and AI-driven healthcare, Fitness and Sports in Health Technology, Mental Health based on Behavior Analysis, and Healthcare Data Privacy with Blockchain. These articles collectively advance our understanding of AI applications in healthcare and reveal significant advancements in diagnostic accuracy, patient monitoring, data security and predictive analytics. Additionally, this review discusses the research findings and implications that will lead to groundbreaking insights in real-world applications, which could provide a clear path for learning a current healthcare technology in mobile health with AI.</p>\u0000 </div>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"7 6","pages":""},"PeriodicalIF":0.9,"publicationDate":"2025-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143187042","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":"Hadamard and Riemann Matrix-Based SLM for PAPR Reduction in OTFS Signal","authors":"Aare Gopal, Desireddy Krishna Reddy","doi":"10.1002/itl2.591","DOIUrl":"https://doi.org/10.1002/itl2.591","url":null,"abstract":"<div>\u0000 \u0000 <p>In this letter, we study the peak-to-average power ratio (PAPR) reduction and bit error rate (BER) performances of the Hadamard and Riemann matrix-based selected mapping (SLM) technique for Orthogonal Time Frequency Space (OTFS) signals. Unlike the conventional phase sequence based on <span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <mrow>\u0000 <mo>{</mo>\u0000 <mrow>\u0000 <mo>±</mo>\u0000 <mn>1</mn>\u0000 </mrow>\u0000 <mo>,</mo>\u0000 <mrow>\u0000 <mo>±</mo>\u0000 <mi>j</mi>\u0000 </mrow>\u0000 <mo>}</mo>\u0000 </mrow>\u0000 </mrow>\u0000 <annotation>$$ left{pm 1,pm jright} $$</annotation>\u0000 </semantics></math>, which requires the entire sequence to extract the original signal at the receiver, the Hadamard and Riemann matrix-based SLM techniques require only the row index of the matrix, reducing the additional information necessary to extract the signal. Simulation results are presented to verify the PAPR and BER performance. The results are also compared with the existing normalized <i>μ</i>-law, rooting <i>μ</i>-law, and conventional SLM methods. The results demonstrate that the Riemann-based SLM technique shows significant performance improvement.</p>\u0000 </div>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"7 6","pages":""},"PeriodicalIF":0.9,"publicationDate":"2024-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143187192","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":"NB-IoT: Analytical Perspective of IP Data Delivery Variants","authors":"Priyatosh Mandal","doi":"10.1002/itl2.594","DOIUrl":"https://doi.org/10.1002/itl2.594","url":null,"abstract":"<div>\u0000 \u0000 <p>Narrowband Internet of Things (NB-IoT) is generally used for transmission of IP data and Non IP data from sensing device to a designated IoT server. Further, the designated server takes decision and accordingly send data to necessary devices for some designated activity. Two communication architectures such as control plane based and user plane based IP data deliveries are possible. In this work, we analytically find the delay to send NB-IoT IP data considering control plane as well as user plane of LTE-A network. Further, we compare the delay considering these two variants of data deliveries. Comparison results show that delay of the control plane based data delivery can be <span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <mn>33</mn>\u0000 <mo>%</mo>\u0000 </mrow>\u0000 <annotation>$$ 33% $$</annotation>\u0000 </semantics></math> more than the delay of user plane based data delivery.</p>\u0000 </div>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"7 6","pages":""},"PeriodicalIF":0.9,"publicationDate":"2024-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143187115","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":"Enhancing Cross-Domain Book Classification Through Caching-Enabled Networks and Transformer Technology","authors":"Qiang Li, Huaiyuan Zheng, Yulai Bao, Side Liu","doi":"10.1002/itl2.590","DOIUrl":"https://doi.org/10.1002/itl2.590","url":null,"abstract":"<div>\u0000 \u0000 <p>Book classification is a crucial task for libraries and a fundamental aspect of their service offerings. Cross-domain book classification, in particular, presents significant challenges due to the diversity and complexity of content across different genres and subjects. To tackle these challenges, a user-oriented strategy employing Transformer network (TN) is developed to fulfill the need for superior content quality and classification. Our proposed method leverages the self-attention mechanism of TN for precise feature extraction and classification, combining it with principal component analysis to ensure a comprehensive understanding of book content. This integration represents a technical innovation that enhances the model's ability to handle diverse datasets with improved accuracy and robustness. Our approach merges TN with caching-enabled networks (CEN) to enhance accuracy and robustness. Driven by the necessity for improved cross-domain classification, our strategy aims to standardize book classifications, thus improving user satisfaction. The primary actions encompass improved classification management, feedback systems, and evaluation frameworks. This work highlights the innovative fusion of TN and CEN, showcasing how these advanced techniques can significantly elevate the performance of library classification systems. Our work demonstrates that high-quality book classification can significantly improve library services and user experience. Furthermore, it aligns with the broader applications of CEN in emerging networking technologies, showing the potential for cutting-edge techniques to revolutionize library services.</p>\u0000 </div>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"8 3","pages":""},"PeriodicalIF":0.9,"publicationDate":"2024-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143688755","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":"PIoT-oriented multi-target recognition of substation infrared images driven by deep learning","authors":"Min Li, Tou Li, Xuan Zhang, Wei Zhang","doi":"10.1002/itl2.573","DOIUrl":"https://doi.org/10.1002/itl2.573","url":null,"abstract":"<p>Substation infrared imaging plays a crucial role in condition monitoring and fault detection of Power Internet of Things (PIoT). However, the accurate and efficient recognition of multiple targets in substation infrared images remains a challenging task. This paper proposes a deep learning-based multi-target recognition framework for substation infrared images in PIoT. This paper presents a method for recognizing various electrical equipment in infrared images of substations using a faster region-based convolutional neural network (Faster RCNN). The optimization of Faster RCNN includes class rectification inspired by non-maximum suppression (NMS), enabling the correction of misclassified equipment parts and enhancing recognition accuracy. The approach combines NMS and class rectification to retain region proposals with optimal recognition performance. Experimental results demonstrate the effectiveness of the proposed method in improving the recognition accuracy of electrical equipment in infrared images.</p>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"8 3","pages":""},"PeriodicalIF":0.9,"publicationDate":"2024-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143688644","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":"IQL-OCDA: An intelligent Q-learning-based for optimal clustering and data-aggregation for wireless sensor networks","authors":"Arwa N. Aledaily","doi":"10.1002/itl2.572","DOIUrl":"https://doi.org/10.1002/itl2.572","url":null,"abstract":"<p>Wireless sensor networks (WSNs) can suffer from low battery life due to the energy consumption of the routing protocol. Small sensor nodes are often difficult to recharge after deployment. In a WSN, data aggregation is generally used to reduce or eliminate data redundancy between nodes in order to save energy. In the proposed algorithm, sensor nodes are deployed in appropriate clusters and cluster heads are elected using Q-learning techniques. Nodes are clustered based on the mean values computed during the clustering phase. Lastly, a performance evaluation and comparison of existing clustering algorithms are performed based on Intelligent Q-learning. The proposed IQL-OCDA model reduces end-to-end delay by 10.11%, increases throughput by 4.15%, and increases network lifetime by 5.1%.</p>","PeriodicalId":100725,"journal":{"name":"Internet Technology Letters","volume":"8 3","pages":""},"PeriodicalIF":0.9,"publicationDate":"2024-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143726782","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}