检索结果-内蒙古大学图书馆

KDTL: knowledge-distilled transfer learning framework for diagnosing mental disorders using EEG spectrograms

Neural Computing and Applications 2024年第30期36卷 18919-18934页

作者： Singh, Shreyash Jadli, Harshit Padma Priya, R. Surya Prasath, V.B. Department of Software Systems School of Computer Science and Engineering Vellore Institute of Technology Vellore India Division of Biomedical Informatics Cincinnati Children’s Hospital Medical Center CincinnatiOH45229 United States Department of Pediatrics University of Cincinnati College of Medicine CincinnatiOH45257 United States Department of Biomedical Informatics College of Medicine University of Cincinnati CincinnatiOH45267 United States Department of Computer Science University of Cincinnati CincinnatiOH45221 United States

Electroencephalography (EEG) is a well-known modality in neuroscience and is widely used in identifying and classifying neurological disorders. This paper investigates how EEG data can be used along with knowledge distillation-based deep learning models to detect mental disorders like epilepsy and sleep disorders. The EEG signals are converted into time–frequency plots using short-time Fourier transforms. Further, we propose a novel methodology for using knowledge distillation-based transfer learning (KDTL). Knowledge distillation is becoming quite prevalent in the machine learning field and is associated with various applications in our work;we propose its use in the detection of mental disorders from EEG spectrograms. We convert the EEGs using short-term Fourier transform to obtain time–frequency representation and apply teacher-student by first training a large teacher model and use knowledge distillation to train a student model. In our experiments, we found that ConvNext teacher and MobileNet student combination obtained better results. Our proposed KDTL approach is tested on two datasets with multiple cases, namely the Bonn and ISRUC datasets and obtain 98% and 95% accuracies, respectively. Further experimental results show that the overall KDTL methodology can obtain high classification accuracy across both datasets in binary and multiclass classifications and proves to be better than multiple prior works. Further, our KDTL approach provides a way to train lightweight models which have a smaller number of trainable parameters and thereby constitute lower training time overall. Our proposed KDTL-based approach obtained accurate results in diagnosing mental disorders from EEG spectrograms. Compared to other related methods, KDTL outperformed across tasks with obtained good results in both binary and multiclass classifications. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Assessing the Viability of AC to DC Microgrid Conversion: A Comprehensive Analysis of Technical and Economic Factors 59

Assessing the Viability of AC to DC Microgrid Conversion: A ...

引用

59th International Universities Power engineering Conference, UPEC 2024

作者： Charalambous, Chrysanthos Polycarpou, Alexis Efthymiou, Venizelos Georghiou, George E. University of Cyprus Pv Technology Laboratory Department Of Electrical and Computer Engineering Cyprus Phaethon Centre of Excellence for Intelligent Efficient and Sustainable Energy Solutions Nicosia Cyprus Frederick University Department of Electrical Computer Engineering and Informatics Nicosia Cyprus

ISBN: (纸本)9798350379730

In the face of global warming and the acceleration of climate changes, urgent and decisive measures must be taken to mitigate these phenomena. This study addresses this urgent need by exploring the transition from traditional AC distribution systems to more efficient DC and hybrid AC/DC systems, recognizing the utmost importance of both technical developments and economic considerations. Through a thorough overview of existing literature and empirical knowledge gathered from the nanogrid of FOSS Research Centre for Sustainable Energy of University of Cyprus, this research clarifies the inherent advantages of DC and AC/DC hybrid systems. In addition, it presents a detailed case study for the comprehensive assessment of the feasibility and effectiveness of these systems. The preliminary economic analysis carried out in this study demonstrates the cost-effectiveness and performance improvements offered by hybrid AC/DC microgrids compared to conventional AC configurations. According to the findings of this work it is obvious that the ROI increases with the increase of DC loads over AC loads until the case where DC loads reach the 75% of the total load (ROI=13.9%). However, even in the case of Full DC microgrid, the ROI is higher by 1% compared to the case of Full AC microgrid thus significantly reducing the payback period. These findings not only underline the urgency of the transition to more efficient and sustainable energy distribution systems, but also provide valuable information for policymakers, industry stakeholders and researchers. By embracing these innovative technologies, a more resilient and environmentally responsive energy infrastructure can be promoted that is in line with the demands of a rapidly changing global climate. © 2024 IEEE.

关键词： Clean energy

来源：评论

学校读者我要写书评

暂无评论

Using Web Technologies to Implement a Modular Integrated System for Augmented Tourism Destinations 3rd

Using Web Technologies to Implement a Modular Integrated Sys...

引用

3rd International Conference on Novel and Intelligent Digital Systems, NiDS 2023

作者： Palamas, Stergios Voutos, Yorghos Zigoura, Paraskevi Mylonas, Phivos Chasanis, Vasileios Department of Informatics Ionian University Corfu Greece iED - Institute of Entrepreneurship Development Larissa Greece Department of Informatics and Computer Engineering University of West Attica Athens Greece

ISBN: (纸本)9783031440960

Identifying and promoting attractions and points of interest to visitors, especially through multimedia and – lately - AR/VR technologies, has been shown to enhance the pre-traveling and on-site overall experience. Throughout this paper, the development of a modular, web-based, generally applicable platform, consisting of interconnected modules and based on open-source technologies is presented. Not long ago, the deployment of engaging, rich-media user experiences required special equipment and native development tools. The present work demonstrates, how recent advancements in web technologies and modern web browsers (e.g., the WebGL), enable the deployment of such cutting-edge services, on mainstream mobile devices, using exclusively well-established web technologies like HTML, CSS and JavaScript. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Web browsers

来源：评论

学校读者我要写书评

暂无评论

EfficientNetB0-based Automated Diabetic Retinopathy Classification in Fundus Images 3

EfficientNetB0-based Automated Diabetic Retinopathy Classifi...

引用

3rd International Conference on Automation, Computing and Renewable Systems, ICACRS 2024

作者： Majaw, Ezekiel Alaric Sundar, G. Naveen Narmadha, D. Thangavel, Senthil Kumar Venkatesan Ajibesin, Adeyemi Abel Department of Computer Science and Engineering Karunya Institute of Technology and Sciences Coimbatore India Department of Computer Science and Engineering Amrita School of Computing Coimbatore India Department of Information Technology Faculty of Informatics and Design Cape Peninsula University of Technology South Africa

ISBN: (纸本)9798331532420

The most common cause of blindness among diabetics is diabetic retinopathy. This disease must be detected in its early stages as delaying treatment can result in permanent blindness. In today’s modern world with the development of technologies, automated techniques have improved the accuracy and efficiency of detecting and classifying this disease. This study suggests a deep learning model utilizing EfficientNetB0 that focuses on designing a neural network structure that requires fewer parameters and computations compared to traditional architectures. Further to help in increasing the accuracy, the image is being processed with methods like CLAHE which is an image processing algorithm that helps in enhancing the contrast while keeping the noise of the images as low as possible. Also, augmentation is being done on the images and SMOTE is applied to handle class imbalance. We collected the APTOS Blindness Detection dataset consisting of retinal fundus images. The accuracy that is being obtained from the proposed methodology is 98.78% indicating the model's effectiveness in detection and classifying between the various stages of diabetic retinopathy. © 2024 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of YOLOv9 and YOLOv10 Algorithms for Urban Safety Improvement 22

Comparative Analysis of YOLOv9 and YOLOv10 Algorithms for Ur...

引用

22nd IEEE International Symposium on Intelligent Systems and informatics, SISY 2024

作者： Glucina, Matko Zigulic, Nikola Frank, Domagoj Lorencin, Ivan Matika, Dario Istrian University of Applied Sciences Department of Engineering Pula Croatia University of Rijeka Faculty of Engineering Rijeka Croatia University North Koprivnica Department of Computer Science and Informatics Croatia Juraj Dobrila University of Pula Faculty of Informatics Pula Croatia University North Department of Electrical Engineering Varazdin Croatia

ISBN: (纸本)9798350385601

Individual safety in urban and city centers is one of the fundamental rights of every person. Today's innovations open up various possibilities for improving the quality of life and safety, but if they are not applied, they carry the danger of an additional increase in crime or greater insecurity in urban centers. This investigation aims to apply insights into the anticipation of illegal actions by various criminal groups that use illegal means, such as weapons, to get property that does not belong to them. In this research, computer vision and advanced strategies such as YOLOv10 and YOLOv9 are used to compare their properties and performance for improving security in urban and populated areas. The best results were achieved by YOLOv9m with a high percentage of mean average precision in the amount of 0.923, while the worst performance was shown by YOLOv10n with a mean average precision of 0.902. Precision and recall for the YOLOv9 and YOLOv10 models were also taken into account and the results are as follows: 0.984.. 0.993.. 0.979.. and 0.996. © 2024 IEEE.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

UNet Based onMulti-Object Segmentation and Convolution Neural Network for Object Recognition

引用

computers, Materials & Continua 2024年第7期80卷 1563-1580页

作者： Nouf Abdullah Almujally Bisma Riaz Chughtai Naif Al Mudawi Abdulwahab Alazeb Asaad Algarni Hamdan A.Alzahrani Jeongmin Park Department of Information Systems College of Computer and Information SciencesPrincess Nourah bint Abdulrahman UniversityP.O.Box 84428Riyadh11671Saudi Arabia Department of Computer Science Air UniversityIslamabad44000Pakistan Department of Computer Science College of Computer Science and Information SystemNajran UniversityNajran55461Saudi Arabia Department of Computer Sciences Faculty of Computing and Information TechnologyNorthern Border UniversityRafha91911Saudi Arabia Information Technology Department College of Computing and InformaticsSaudi Electronic UniversityRiyadh13316Saudi Arabia Department of Computer Engineering Korea Polytechnic UniversitySiheung-siGyeonggi-do237Republic of Korea

The recent advancements in vision technology have had a significant impact on our ability to identify multiple objects and understand complex *** technologies,such as augmented reality-driven scene integration,robotic navigation,autonomous driving,and guided tour systems,heavily rely on this type of scene *** paper presents a novel segmentation approach based on the UNet network model,aimed at recognizing multiple objects within an *** methodology begins with the acquisition and preprocessing of the image,followed by segmentation using the fine-tuned UNet ***,we use an annotation tool to accurately label the segmented *** labeling,significant features are extracted from these segmented objects,encompassing KAZE(Accelerated Segmentation and Extraction)features,energy-based edge detection,frequency-based,and blob *** the classification stage,a convolution neural network(CNN)is *** comprehensive methodology demonstrates a robust framework for achieving accurate and efficient recognition of multiple objects in *** experimental results,which include complex object datasets like MSRC-v2 and PASCAL-VOC12,have been *** analyzing the experimental results,it was found that the PASCAL-VOC12 dataset achieved an accuracy rate of 95%,while the MSRC-v2 dataset achieved an accuracy of 89%.The evaluation performed on these diverse datasets highlights a notably impressive level of performance.

关键词： UNet segmentation blob fourier transform convolution neural network

来源：评论

学校读者我要写书评

暂无评论

Relevance Vector Machine for Code Smell Detection 22

Relevance Vector Machine for Code Smell Detection

引用

22nd IEEE International Conference on Industrial informatics, INDIN 2024

作者： Putro, Hanson Prihantoro Yuhana, Umi Laili Yuniarno, Eko Mulyanto Purnomo, Mauridhi Hery Department of Electrical Engineering Institut Teknologi Sepuluh Nopember Surabaya Indonesia Department of Informatics Institut Teknologi Sepuluh Nopember Surabaya Indonesia Department of Computer Engineering Department of Electrical Engineering Institut Teknologi Sepuluh Nopember Surabaya Indonesia

ISBN: (纸本)9798331527471

Various approaches have been proposed to detect code smells, including machine learning models. However, there are still challenges to improving detection accuracy and selecting appropriate quality metrics. This research introduces the use of Relevance Vector Machine (RVM) for code smell detection, especially Magic Value smells, which often affect source code readability and maintainability in a software development industry. This approach utilizes source code analysis using two source code representations (raw source code and token stream) and feature extraction using two text mining techniques (Bag of Words/BoW and Term Frequency-Inverse Document Frequency/TF-IDF). The experiment involves analyzing 983 records of labeled source code datasets. The results show that RVM can detect code smells with a very high level of accuracy, reaching 99.9% on some kernel configurations, also nearly perfect precision, and F1 scores. This paper contributes to introducing the use of RVM for detecting code smells in source code, explores source code representation and feature extraction techniques, and shows that RVM can achieve high detection accuracy. © 2024 IEEE.

关键词： Software design

来源：评论

学校读者我要写书评

暂无评论

Application of Machine Learning Models to Predict e-Learning Engagement Using EEG Data 20

Application of Machine Learning Models to Predict e-Learning...

引用

20th International Conference on Web Information Systems and Technologies, WEBIST 2024

作者： Dritsas, Elias Trigka, Maria Mylonas, Phivos Industrial Systems Institute Athena Research and Innovation Center Patras Greece Department of Informatics and Computer Engineering University of West Attica Greece

ISBN: (纸本)9789897587184

The rapid evolution of e-learning platforms necessitates the development of innovative methods to enhance learner engagement. This study leverages machine learning (ML) techniques and models to predict e-learning engagement with the aid of Electroencephalography (EEG). Various ML models, including Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting Machine (GBM), and Neural Networks (NN), were applied to a dataset comprising EEG signals collected during e-learning sessions. Among these models, NN demonstrated the highest accuracy (90%), with precision and F1-score of 88%, a recall of 89%, and an Area Under the Curve (AUC) of 0.92 for predicting engagement levels. The results underscore the potential of EEG-based analysis combined with advanced ML techniques to optimize e-learning environments by accurately monitoring and responding to learner engagement. Copyright © 2024 by SCITEPRESS.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

YOLOv8 vs. YOLOv9: Evaluating Object Detection Algorithms for Marine Waste Recognition 1

YOLOv8 vs. YOLOv9: Evaluating Object Detection Algorithms fo...

引用

1st IEEE International Symposium on Consumer Technology, ISCT 2024

作者： Wajdi, Achmad Zahir Wibowo, Prasetyo Sasaki, Shiori Politeknik Elektronika Negeri Surabaya Department of Informatics and Computer Engineering Surabaya Indonesia Faculty of Data Science Musashino University Tokyo Japan

ISBN: (纸本)9798350365191

Marine debris, the majority of which is composed of plastic (61 % to 87 %), is a significant environmental issue facing the world. Between 4.8 million and 12.7 million metric tons of plastic are thought to have entered the ocean in 2010 alone, increasing the estimated total amount of particles in the water to at least 5 trillion. Global plastic production, which reached 348 million tons in 2017, is expected to double in the next 20 years. To overcome this challenge, effective waste detection methods are needed, one of which is by utilizing computer vision, specifically the You Only Look Once (YOLO) algorithm. YOLOv9, as the latest version, offers improved object detection performance compared to its predecessor, YOLOv8. This research aims to evaluate the performance and efficiency of YOLOv9 and YOLOv8 in detecting marine waste. Using the Trash Annotations in Context dataset (TACO), We train both models and conduct tests to evaluate their accuracy and efficiency. Experimental results show that YOLOv9 has higher precision at 45% but lower recall at 6%, compared to YOLOv8, which has a precision of 33 % and a recall of 15 %. Overall, YOLOv8 shows better performance with higher mAP50 at 13% and mAP50-95 values at 10%, compared to YOLOv9, which has mAP50 at 4% and mAP50-95 values at 3%. © 2024 IEEE.

关键词： Marine Waste Neural Net-work Object Detection TACO Dataset YOLO

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Ensemble Deep Learning Approach for Emotion Classification

A Hybrid Ensemble Deep Learning Approach for Emotion Classif...

引用

2022 IEEE International Conference on Big Data, Big Data 2022

作者： Karras, Christos Karras, Aristeidis Tsolis, Dimitrios Avlonitis, Markos Sioutas, Spyros Group Computer Engineering and Informatics Department Patras Greece Ionian University Department of Informatics Corfu Greece

ISBN: (纸本)9781665480451

Speech processing, the field of analysing input speech signals and methods of processing them has emerged in the recent days. Additionally, the development of a speech processing system involves several components in the design phase with probabilistic approximations for enhanced audio sampling and de-noising. In this work, we focus into use of Gaussian random variables while modelling and filtering noise that gets added after being passed through an additive noise channel in a communication system, and the applications of Hidden Markov models. Moreover, we apply deep learning methods for emotion classification via a robust and accurate ensemble learning scheme that is applied to a joint deep network which incorporates audiovisual inputs and generates the emotion prediction effectively reaching satisfactory accuracy. © 2022 IEEE.

关键词： Emotion Recognition

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：