检索结果-内蒙古大学图书馆

Abnormal Activity Detection and Classification of Bus Passengers with In-Vehicle Image Sensing

IEEE Access 2024年 12卷 23057-23065页

作者： Lin, Huei-Yung Tseng, Chun-Han National Taipei University of Technology Department of Computer Science and Information Engineering Taipei106 Taiwan National Chung Cheng University Department of Electrical Engineering Chiayi621 Taiwan

As the self-driving technology is getting mature for public transportation applications, the safety concern of onboard passengers has become an important issue. It is essential to identify inappropriate or hazardous behaviors of passengers for the vehicles without human operators. In this work, we propose a technique to detect and classify the abnormal activities of passengers in a bus environment. Different from the existing human activity classification algorithms, our approach reduces the occlusion and increases the recognition rate by acquiring images from an overhead vision system. To overcome the increased complexity on feature extraction and classification, an action recognition network for top-view images are proposed by incorporating both spatial and temporal information. An image dataset, BUS-HAR, is generated for practical application scenarios with bus passengers. Experiments using real-world scene images have demonstrated the feasibility of our technique compared to existing approaches. The codes and image dataset are made available publicly at https://***/richardkuo1999/passenger-action-recognition. © 2013 IEEE.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

A Grid-Based Random Erasing Approach for Enhanced Image Classification 4

A Grid-Based Random Erasing Approach for Enhanced Image Clas...

引用

4th IEEE International Conference on Data science and computer Application, ICDSCA 2024

作者： Li, Qingyi School of Electrical Engineering and Computer Science The University of Queensland Brisbane Australia

ISBN: (纸本)9798350368239

In modern image classification tasks, the data is insufficient due to time or money cost constraints, which affects the generalization and accuracy of the model. This paper proposes a random erasure method based on gridded images, aiming to expand the dataset to improve the generalization ability and classification accuracy of the model while reducing the damage to the original structure of the image caused by continuous erasure. This paper uses YOLOv8n-cls, YOLOv8mcls, and ResNet50 as models and selects CIFAR-10, CIFAR-100, FER2013, and mini-ImageNet as datasets for comparative experiments. The Top-1 accuracy, Top-5 accuracy, and loss function curves on the validation set are used to analyze the effect of random erasing on each model's performance on various datasets. The findings demonstrate that lightweight models greatly increase accuracy on simple datasets like CIFAR-10. However, to effectively use the enhancing effect of random erasing on complicated datasets, more sophisticated models are required. This study shows the impact of random erasing based on gridded images on improving the generalization ability and classification accuracy of the model in image classification tasks. It points out the limitations of this method and potential future optimization directions to better handle more complex classification tasks. ©2024 IEEE.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

An Intrusion Detection and Mitigation Framework for Automatic Generation Control Systems

IEEE Transactions on Industrial Cyber-Physical Systems

引用

IEEE Transactions on Industrial Cyber-Physical Systems 2024年 2卷 412-421页

作者： Mohammadi, Fazel Saif, Mehrdad University of New Haven Electrical and Computer Engineering and Computer Science Department West HavenCT06516 United States University of Windsor Department of Electrical and Computer Engineering WindsorONN9B 3P4 Canada

The main role of Automatic Generation Control (AGC) is to maintain power grids frequency within specified operating limits. Due to the fact that AGC is the sole automatic feedback control loop between physical and cyber infrastructure in modern power systems and the data required by the AGC system is transferred to a control center through communication links, it can be highly vulnerable to malicious attacks. Therefore, AGC systems should be well-protected against cyberattacks, e.g., False Data Injection (FDI) attacks. In this paper, an intrusion detection and mitigation framework for AGC systems based on a modified Goertzel algorithm is proposed. Compared with the existing intrusion detection and mitigation strategies, the major superiorities of the proposed framework are less computational burden, high accuracy, and rapid detection and mitigation of FDI attacks, which are considered unknown inputs. The proposed framework is validated on a two-area interconnected power systems model and the IEEE 39-bus test system. The dynamic simulation results under different testing conditions verify the applicability and effectiveness of the proposed framework. © 2023 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Graph Collaborative Filtering and Data Augmentation Strategies in Dual-Target CDR 29th

Graph Collaborative Filtering and Data Augmentation Strateg...

引用

29th International Conference on Cooperative Information Systems, CoopIS 2023

作者： Shao, Xiaowen Liu, Baisong Zhang, Xueyuan Li, Junru Xu, Ercong Wu, Shiqi Faculty of Electrical Engineering and Computer Science NingBo University Ningbo China

ISBN: (纸本)9783031468452

Current dual-target recommendation methods focus on efficient feature fusion but neglect the inherent noise issues in the domains. However, noise negatively affects the fusion of domains. To tackle this issue, we introduce an improvement and noise reduction strategy named DA-DCDR(Data Augmentation-Dual Target Cross Domain Recommendation), for domain fusion. By refining and reducing noise in each domain’s subgraphs, we not only enhance the accuracy of interaction data but also ensure consistency in data scales. To establish associations between distinct domains, we implement a graph co-training strategy. Key procedures of DA-DCDR include interaction refinement and noise reduction, domain fusion, and correlation expansion. We use graph encoders to acquire user/item embeddings for both domains before domain fusion, followed by enhancement and noise reduction in interactions via top-k sampling and re-prediction. Additionally, we amplify user-user and item-item correlation elements after the domain fusion. Experimental results validate the noteworthy performance enhancement of our proposed strategy in the dual-target recommendation, mitigating the noise effects and boosting the accuracy of the dual-target recommendation system. © 2024, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

Automated Fact Checking Using A Knowledge Graph-based Model 6

Automated Fact Checking Using A Knowledge Graph-based Model

引用

6th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2024

作者： Kundu, Arghya Nguyen, Uyen Trang York University Department of Electrical Engineering and Computer Science Toronto Canada

ISBN: (纸本)9798350344349

Misinformation is a growing threat to the economy, social stability, public health, democracy, and national security. One of the most effective methods to combat misinformation is fact checking. Fact checking is the process of verifying the factual accuracy of a statement or claim. Fact checkers employ rigorous methodologies to scrutinize claims, verify sources, and expose falsehoods. However, the huge volume of content circulating online makes it challenging for humans to identify misinformation manually. Automated tools can analyze large datasets to detect patterns in misinformation content, scaling up fact checking efforts. This paper proposes a knowledge graph-based fact checking model that uses two separate knowledge graphs, one containing true claims and the other, false claims. The model uses knowledge graph embeddings which are based on convolutional neural networks. The deep learning model is trained on the above two knowledge graphs to learn distinguishing patterns between true and false claims. Additionally, we employ explainable artificial intelligence (XAI) techniques to provide explanations for the model's classification, reducing cost of errors and increasing transparency and user trust in the system. © 2024 IEEE.

关键词： Large datasets

来源：评论

学校读者我要写书评

暂无评论

A Sub-nanosecond Delay Floating-voltage Level Shifter for SiC MOSFETs Gate Driver Application 6

A Sub-nanosecond Delay Floating-voltage Level Shifter for Si...

引用

6th International Conference on Energy Systems and electrical Power, ICESEP 2024

作者： Wan, Yue Xia, Yinshui Ningbo University Faculty of Electrical Engineering and Computer Science Ningbo China

ISBN: (纸本)9798350353563

In order to meet the requirements for the design of SiC gate driver in terms of speed and dv/dt immunity, a sub-nanosecond delay level shifter based on a high-voltage (HV) digital level shifter circuit is proposed. It can decrease the response time compared with the existing level-shifting circuitry, enabling its application in high-voltage scenarios such as closed-loop drive circuits for silicon carbide. The edge detection circuit and simple pull-up branches are utilized to reduce the transmission delay and power consumption. The circuit is simulated at a frequency of 10MHz in a 018um BCD process with Virtuoso, and the results demonstrate that the average transmission delay is about 380ps and the charging response time of the slow path is reduced to 3.3ns. © 2024 IEEE.

关键词： Delay circuits

来源：评论

学校读者我要写书评

暂无评论

Detection of Novel Cyberattacks Using Multi-Binary Classifiers 34

Detection of Novel Cyberattacks Using Multi-Binary Classifie...

引用

34th Annual International Conference on Collaborative Advances in Software and Computing, CASCON 2024

作者： Gaudreault, Jean-Gabriel Branco, Paula School of Electrical Engineering and Computer Science University of Ottawa OttawaON Canada

ISBN: (纸本)9798331504830

Intrusion Detection Systems (IDSs) are an essential part of modern-day cybersecurity defenses, enabling the detection of potential malicious activities within network and computer systems. Although multiple techniques have been employed to implement IDSs, supervised machine learning approaches have historically proven particularly effective due to their capacity to leverage large amounts of data and maintain a generally low false alarm rate. However, a significant limitation of these approaches is their inability to detect novel threats, such as zero-day attacks, that were not present in the training data. In this study, we introduce a novel supervised approach based on binary classifiers that can detect these previously unseen attacks while also maintaining a high detection rate and low false alarm rate through the use of a filtering mechanism. We empirically demonstrate the effectiveness of our proposed technique on multiple intrusion detection datasets, showing that it is able to detect unseen attacks and perform multiclass classification while outperforming traditional supervised methods. © 2024 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Real-Time Spectrum Sensing for TDD-LTE Signals Using Deep Learning 16

Real-Time Spectrum Sensing for TDD-LTE Signals Using Deep Le...

引用

16th International Conference on Communication Software and Networks, ICCSN 2024

作者： Xu, Xiang Jin, Ming Fang, Xiangqian Jiang, Tao Yu, Yanghang Li, Yaming Ningbo University Faculty of Electrical Engineering and Computer Science Ningbo China

ISBN: (纸本)9798331506322

Existing spectrum sensing techniques for Time Division Duplex-Long Term Evolution (TDD-LTE) signals primarily focus on large time-scale signals. However, the spectrum holes that exist in the small time-scale special subframes of TDD-LTE are often ignored. This paper proposes a method for spectrum sensing within TDD-LTE special subframes by applying Parallel Convolutional Neural Networks (ParallelCNN). The effective feature extraction capability of ParallelCNN enhances the spectrum sensing performance. Simulation results demonstrate superior sensing performance under various signal-to-noise ratio (SNR) conditions. © 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

A Wait-Free Deque With Polylogarithmic Step Complexity 27

A Wait-Free Deque With Polylogarithmic Step Complexity

引用

27th International Conference on Principles of Distributed Systems, OPODIS 2023

作者： Asbell, Shalom M. Ruppert, Eric Department of Electrical Engineering and Computer Science York University Toronto Canada

ISBN: (纸本)9783959773089

The amortized step complexity of operations on all previous lock-free implementations of double-ended queues is linear in the number of processes. This paper presents the first concurrent double-ended queue where the amortized step complexity of each operation is polylogarithmic. Since a stack is a special case of a double-ended queue, this is also the first concurrent stack with polylogarithmic step complexity. The implementation is wait-free and the amortized step complexity is O(log2 p + log q) per operation, where p is the number of processes and q is the size of the double-ended queue. © Shalom M. Asbell and Eric Ruppert

关键词： Queueing theory

来源：评论

学校读者我要写书评

暂无评论

Millimeter-Wave Radar-Based Vital Sign Signal Extraction Using Multiple Distance Bin

Millimeter-Wave Radar-Based Vital Sign Signal Extraction Usi...

引用

2024 International Applied Computational Electromagnetics Society Symposium, ACES-China 2024

作者： Zheng, Zhiwei Huo, Jianjian Wang, Jian Ningbo University Faculty of Electrical Engineering and Computer Science Ningbo China

ISBN: (纸本)9798350355581

This paper introduces a multi-distance non-contact vital sign signal reconstruction and extraction algorithm based on frequency modulated continuous wave (FMCW) millimeter wave radar. The technology of measuring vital signs with millimeter wave radar has significant medical value. However, issues such as environmental noise, unconscious micro-movements of the human body, and the extremely subtle vibrations of the thoracic cavity caused by heartbeats greatly reduce the signal-to-noise ratio of the vital sign signals. To address these issues, considering the characteristic that the human body is perceived as a point cloud target under millimeter wave radar, this paper proposes a technique to enhance the signal-to-noise ratio of weak heartbeat signals by utilizing multi-distance vital sign signals. Furthermore, this paper employs the variational mode decomposition (VMD) algorithm for the reconstruction of vital sign signals, achieving high-precision extraction of the human heart rate signal. © 2024 IEEE.

关键词： Signal to noise ratio

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：