检索结果-内蒙古大学图书馆

MobileDNN: Leveraging the Potential of Hybrid MobileNetV2-DNN for Advanced Image Analysis

MobileDNN: Leveraging the Potential of Hybrid MobileNetV2-DN...

2024 International Conference on Science, engineering and Business for Driving Sustainable Development Goals, SEB4SDG 2024

作者： Ogundokun, Roseline Oluwaseun Maskeliunas, Rytis Damasevicius, Robertas Adebiyi, Marion Olubunmi Landmark University Department of Computer Science Omu Aran Nigeria Kaunas University of Technology Department of Multimedia Engineering Lithuania Kaunas University of Technology Faculty of Informatics Kaunas Lithuania

ISBN: (纸本)9798350358155

computer vision systems designed for posture recognition could offer valuable solutions in the healthcare sector, improving healthy ageing and supporting elderly individuals in their daily routines. One area of interest is the automatic detection of falls, which has garnered attention from computer vision and pattern recognition communities. While many existing methods rely on wearable sensors, which demonstrate high accuracy in fall detection, some individuals may be reluctant to wear such devices. Consequently, alternative approaches, such as vision-based techniques, have been developed. The proposed method utilizes a vision-based system that employs a convolutional neural network (CNN) to recognize and categorize various postures (e.g., sitting, lying, standing) in each frame, if available. This approach begins by detecting the person in the frame, applying background subtraction for human body shape extraction, and associating each daily activity with the extracted shape. The proposed model achieved an accuracy of 93.75%, which shows an outstanding performance over existing models. Subsequently, it estimates bounding boxes encompassing the predicted person's body shape. Overall, the proposed technique shows promising results for indoor monitoring systems. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

PToPI: A Comprehensive Review, Analysis, and Knowledge Representation of Binary Classification Performance Measures/Metrics

引用

SN computer Science 2023年第1期4卷 13页

作者： Canbek, Gürol Taskaya Temizel, Tugba Sagiroglu, Seref Pointr Ankara Turkey Informatics Institute Middle East Technical University Ankara Turkey Computer Engineering Department Gazi University Ankara Turkey

Although few performance evaluation instruments have been used conventionally in different machine learning-based classification problem domains, there are numerous ones defined in the literature. This study reviews and describes performance instruments via formally defined novel concepts and clarifies the terminology. The study first highlights the issues in performance evaluation via a survey of 78 mobile-malware classification studies and reviews terminology. Based on three research questions, it proposes novel concepts to identify characteristics, similarities, and differences of instruments that are categorized into ‘performance measures’ and ‘performance metrics’ in the classification context for the first time. The concepts reflecting the intrinsic properties of instruments such as canonical form, geometry, duality, complementation, dependency, and leveling, aim to reveal similarities and differences of numerous instruments, such as redundancy and ground-truth versus prediction focuses. As an application of knowledge representation, we introduced a new exploratory table called PToPI (Periodic Table of Performance Instruments) for 29 measures and 28 metrics (69 instruments including variant and parametric ones). Visualizing proposed concepts, PToPI provides a new relational structure for the instruments including graphical, probabilistic, and entropic ones to see their properties and dependencies all in one place. Applications of the exploratory table in six examples from different domains in the literature have shown that PToPI aids overall instrument analysis and selection of the proper performance metrics according to the specific requirements of a classification problem. We expect that the proposed concepts and PToPI will help researchers comprehend and use the instruments and follow a systematic approach to classification performance evaluation and publication. © 2022, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Classification Knowledge representation Machine learning Performance evaluation Performance measures Performance metrics Periodic table

来源：评论

学校读者我要写书评

暂无评论

A Double Threshold Energy Detection-Based Neural Network for Cognitive Radio Networks

引用

computer Systems Science & engineering 2023年第4期45卷 329-342页

作者： Nada M.Elfatih Elmustafa Sayed Ali Maha Abdelhaq Raed Alsaqour Rashid A.Saeed Department of Electrical and Electronics Engineering Red Sea UniversityPort Sudan33311Sudan Department of Information Technology College of Computer and Information SciencesPrincess Nourah bint Abdulrahman UniversityRiyadh11671Saudi Arabia Department of Information Technology College of Computing and InformaticsSaudi Electronic UniversityRiyadh93499Saudi Arabia Department of Computer Engineering College of Computers and Information TechnologyTaif UniversityTaif21944Saudi Arabia Department of Electronics Engineering Sudan University of Science and TechnologyKhartoum11111Sudan

In cognitive radio networks(CoR),the performance of cooperative spectrum sensing is improved by reducing the overall error rate or maximizing the detection *** optimization methods are usually used to optimize the number of user-chosen for cooperation and the threshold ***,these methods do not take into account the effect of sample size and its effect on improving CoR *** general,a large sample size results in more reliable detection,but takes longer sensing time and increases ***,the locally sensed sample size is an optimization ***,optimizing the local sample size for each cognitive user helps to improve CoR *** this study,two new methods are proposed to find the optimum sample size to achieve objective-based improved(single/double)threshold energy detection,these methods are the optimum sample size N^(*)and neural networks(NN)*** the evaluation,it was found that the proposed methods outperform the traditional sample size selection in terms of the total error rate,detection probability,and throughput.

关键词： Cognitive radio spectrum sensing energy detection double threshold neural network machine learning optimization quality of service

来源：评论

学校读者我要写书评

暂无评论

Roller Vibration Detection with Radar Sensor and Linear Discriminant Analysis 26

Roller Vibration Detection with Radar Sensor and Linear Disc...

引用

26th International Conference on Digital Signal Processing and its Applications, DSPA 2024

作者： Valuyskiy, Denis Vityazev, Sergey Vashkevich, Maxim Ryazan State Radio Engineering University Named after V.F. Utkin Ryazan Russia University of Informatics and Radioelectronics Computer Engineering Department of Belarussian State Minsk Belarus

ISBN: (纸本)9798350361728

This article is dedicated to research on the problems of quality control of rotation of industrial rollers using radar equipment. The radar signals reflected from roller are analyzed using short-time Fourier transform and classified using linear discriminant analysis technique. To evaluate the proposed approach an testbed has been implemented. To obtain a dataset a set of radar signals has been recorded. The results of experiments shown that the accuracy of the proposed approach of detecting roller condition is reach 98%. © 2024 IEEE.

关键词： Rollers (machine components)

来源：评论

学校读者我要写书评

暂无评论

K-Banhatti Invariants Empowered Topological Investigation of Bridge Networks

引用

computers, Materials & Continua 2022年第12期73卷 5423-5440页

作者： Khalid Hamid Muhammad Waseem Iqbal Erssa Arif Yasir Mahmood Ahmad Salman Khan Nazri Kama Azri Azmi Atif Ikram Department of Computer Science Superior UniversityLahore54000Pakistan Department of Software Engineering Superior UniversityLahore54000Pakistan Department of Software Engineering Faculty of Information TechnologyThe University of Lahore54000Pakistan Advanced Informatics Department Razak Faculty of Technology and InformaticsUniversiti Teknologi MalaysiaKuala Lumpur54100Malaysia Faculty of Informatics and Applied Mathematics University Malaysia TerengganuMalaysia Department of Computer Science&Information Technology Faculty of Information TechnologyThe University of Lahore54000Pakistan

Any number that can be uniquely determined by a graph is called graph *** the most recent twenty years’innumerable numerical graph invariants have been described and used for correlation *** the fast and advanced environment of manufacturing of networks and other products which used different networks,no dependable assessment has been embraced to choose,how much these invariants are connected with a network graph or molecular *** this paper,it will talk about three distinct variations of bridge networks with great capability of expectation in the field of computer science,chemistry,physics,drug industry,informatics,and mathematics in setting with physical and synthetic constructions and networks,since K-Banhatti invariants are newly introduced and have various forecast characteristics for various variations of bridge graphs or *** review settled the topology of bridge graph/networks of three unique sorts with three types of K-Banhatti *** concluded outcomes can be utilized for the modeling of interconnection networks of Personal computers(PC),networks like Local area network(LAN),Metropolitan area network(MAN)and Wide area network(WAN),the spine of internet and different networks/designs of PCs,power generation interconnection,bio-informatics and chemical structures.

关键词： Bridge networks invariants k-banhatti indices maple network graph molecular graph

来源：评论

学校读者我要写书评

暂无评论

Adjusting Logit in Gaussian Form for Long-Tailed Visual Recognition

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第10期5卷 5026-5039页

作者： Li, Mengke Cheung, Yiu-Ming Lu, Yang Hu, Zhikai Lan, Weichao Huang, Hui Shenzhen518132 China Hong Kong Baptist University Department of Computer Science 999077 Hong Kong Xiamen University Fujian Key Laboratory of Sensing and Computing for Smart City School of Informatics Xiamen361005 China Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China

It is not uncommon that real-world data are distributed with a long tail. For such data, the learning of deep neural networks becomes challenging because it is hard to classify tail classes correctly. In the literature, several existing methods have addressed this problem by reducing classifier bias, provided that the features obtained with long-tailed data are representative enough. However, we find that training directly on long-tailed data leads to uneven embedding space. That is, the embedding space of head classes severely compresses that of tail classes, which is not conducive to subsequent classifier learning. This article therefore studies the problem of long-tailed visual recognition from the perspective of feature level. We introduce feature augmentation to balance the embedding distribution. The features of different classes are perturbed with varying amplitudes in Gaussian form. Based on these perturbed features, two novel logit adjustment methods are proposed to improve model performance at a modest computational overhead. Subsequently, the distorted embedding spaces of all classes can be calibrated. In such balanced-distributed embedding spaces, the biased classifier can be eliminated by simply retraining the classifier with class-balanced sampling data. Extensive experiments conducted on benchmark datasets demonstrate the superior performance of the proposed method over the state-of-the-art ones. © 2020 IEEE.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Hybrid Local Cache Model for Enhancing Performance and Memory Footprint of Mobile Heap Allocator 12

Hybrid Local Cache Model for Enhancing Performance and Memor...

引用

12th IEEE Global Conference on Consumer Electronics, GCCE 2023

作者： Lee, Kyung-Seok Lee, Teahyung Eom, Young Ik Sungkyunkwan University Department of Digital Media and Communications Engineering Suwon Korea Republic of Sungkyunkwan University Department of Computer Science and Engineering Suwon Korea Republic of Sungkyunkwan University College of Computing and Informatics Department of Electrical and Computer Engineering Suwon Korea Republic of

ISBN: (纸本)9798350340181

The representative mobile heap allocator that has been used in Android system, called SCUDO, supports two local cache models: the shared model and the exclusive model. However, neither the shared model nor the exclusive model cannot satisfy both memory usage and performance requirements simultaneously since mobile applications have diverse memory allocation/free patterns and the number of threads varies from single to several tens. This paper suggests a hybrid local cache model that aims to adaptively meet memory usage and performance goals. Our evaluation results show that our hybrid local cache model outperforms the shared model by up to 22% on allocation time, while preserving the memory usage. © 2023 IEEE.

关键词： Android (operating system)

来源：评论

学校读者我要写书评

暂无评论

Total Labelings of Graphs with Prescribed Weights

引用

Journal of Combinatorial Mathematics and Combinatorial Computing 2024年 120卷 367-376页

作者： Bača, Martin Miller, Mirka Phanalasy, Oudone Ryan, Joe Semaničová-Feňovčíková, Andrea Sillasen, Anita A. Department of Applied Mathematics and Informatics Technical University Košice Slovakia School of Mathematical and Physical Sciences The University of Newcastle Australia School of Electrical Engineering and Computer Science The University of Newcastle Australia Department of Mathematics University of West Bohemia Pilsen Czech Republic Department of Informatics King's College London United Kingdom Department of Mathematics National University of Laos Vientiane Laos Department of Mathematical Sciences Aalborg University Aalborg Denmark

The total labeling of a graph G = (V, E) is a bijection from the union of the vertex set and the edge set of G to the set {1, 2, . . . , |V(G)| + |E(G)|}. The edge-weight of an edge under a total labeling is the sum of the label of the edge and the labels of the end vertices of that edge. The vertexweight of a vertex under a total labeling is the sum of the label of the vertex and the labels of all the edges incident with that vertex. A total labeling is called edge-magic or vertex-magic when all the edge-weights or all the vertex-weights are the same, respectively. When all the edge-weights or all the vertex-weights are different then a total labeling is called edge-antimagic or vertex-antimagic total, respectively. In this paper we deal with the problem of finding a total labeling of some classes of graphs that is simultaneously vertex-magic and edge-antimagic or simultaneously vertex-antimagic and edge-magic, respectively. We show several results for stars, paths and cycles. © 2024 Charles Babbage Research Centre. All rights reserved.

关键词： Edge-antimagic total labeling Edge-magic total labeling Vertex-antimagic total labeling Vertex-magic total labeling

来源：评论

学校读者我要写书评

暂无评论

Privacy-Preserving Collaborative AI for Distributed Deep Learning with Cross-sectional Data 5

Privacy-Preserving Collaborative AI for Distributed Deep Lea...

引用

5th International Conference on Bio-engineering for Smart Technologies, BioSMART 2023

作者： Iqbal, Saeed Qureshi, Adnan N. Aurangzeb, Khursheed Javeed, Khalid University of Central Punjab Faculty of Information Technology & Computer Science Department of Computer Science Lahore Pakistan King Saud University College of Computer and Information Sciences Department of Computer Engineering Riyadh Saudi Arabia University of Sharjah College of Computing and Informatics Department of Computer Engineering Sharjah United Arab Emirates

ISBN: (纸本)9798350338492

The article discusses the challenges of using deep learning in healthcare due to the lack of extensive medical datasets and concerns about confidentiality and privacy. The article then focuses on the analysis of skin lesion images, which are complex and difficult to classify. The article presents a unique CNN model (skin-net) using a progressively private federated learning architecture to potentially address these issues. Recent advancements in medical image analysis using convolutional neural networks and federated learning techniques have shown promising results (96.19 accuracy, 96.37 sensitivity and 96.53 specificity) in accurately classifying skin diseases while maintaining data privacy. Using these methods offers a potentially effective way to examine human skin while improving the model's generalization performance through image augmentation. © 2023 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Improving Information Freshness in Edge-Assisted Smart Grids: An AoI-Aware Routing Strategy for Neighborhood Area Networks 22

Improving Information Freshness in Edge-Assisted Smart Grids...

引用

22nd IEEE International Conference on Industrial informatics, INDIN 2024

作者： Farag, Hossam Kotb, Mostafa Stefanović, Čedomir Gidlund, Mikael Department of Electronic Systems Aalborg University Denmark Department of Informatics Universität Hamburg Germany Department of Electrical Engineering Aswan University Egypt Mathematics Department Aswan University Egypt Department of Computer and Electrical Engineering Mid Sweden University Sweden

ISBN: (纸本)9798331527471

The integration of edge controllers into smart grid infrastructures facilitates advanced functionalities and high responsiveness, thereby bolstering the overall efficiency of the energy grid. The freshness of the sensing information received at the edge controller, captured by the Age of Information (AoI) metric, is vital to maintain system stability, where outdated information may lead to incorrect responses to grid conditions, potentially causing inefficiencies or system disruptions. However, the timeliness of the transmitted updates is mainly compromised by the delay and congestion within the routing links in the Neighborhood Area Network (NAN). In this work, we develop an intelligent routing strategy to improve the AoI of the Routing Protocol for Low-power and lossy networks (RPL), which is the common routing protocol for smart grids. Our proposed method is based on an AoI-aware parent selection mechanism, by which a node becomes attached to the parent with the highest probability of delivering a packet within a predefined AoI threshold. The prediction is made based on a supervised machine learning model, trained using the collection of heterogeneous routing metrics. The performance of the proposed method is evaluated via extensive discrete-event simulations and the results show its potency to improve the peak AoI and the AoI violation probability compared to the standard RPL. © 2024 IEEE.

关键词： Edge computing

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：