检索结果-内蒙古大学图书馆

2nd International Conference on Machine Learning and Autonomous Systems, ICMLAS 2025

作者： Mahesh Prasanna, K. Marotka, Devashree Muralidharan, J. Adhvaryu, Rachit Vani, V. Maranan, Ramya Vivekananda College of Engineering and Technology Department of Electronics and Communication Engineering Puttur India G H Raisoni College of Engineering and Management Department of Electronics and Telecommunication Engineering Maharashtra Nagpur India KPR Institute of Engineering and Technology Department of Electronics and Communication Engineering Tamil Nadu Coimbatore India Parul institute of Engineering and Technology Parul University Computer science and Engineering Faculty of Engineering and Technology Gujarat India Bangalore Institute of Technology Department of Information Science and Engineering Karnataka Bangalore India Saveetha School of Engineering SIMATS Department of Research and Innovation Tamil Nadu Chennai India

ISBN: (纸本)9798331505745

Internet of Things and Industrial Internet of Things have seen rapid adoption and have transformed industries through smart automation, constant monitoring and better connected systems. The progressive inclusion of such solutions has contributed to the growth in exposure to complex and changing global threats. The primary challenge lies in accurately detecting and classifying cyber threats from high-volume, unstructured network traffic data in real-time. Existing methods are not very accurate and cannot be easily scaled due to constraints in computational resources and the nature of IIoT network systems. This paper introduces Traffic Surveillance System through Capsule Gated Graph Attention Network for IoT/IIoT Cyberthreat Detection and Mitigation (Cyber-CapG). The system uses raw traffic data from Edge-IIoTset and uses the Tunable Transformer for feature extraction. The extracted features are passed through the Capsule Gated Graph Attention Network to model the traffic as a graph based on communication hierarchy and being selective to essential traffic patterns. Furthermore, weights and loss functions are tuned based on the Progressive Crested Porcupine Optimization so as to achieve maximum optimization. The proposed system exhibits outstanding performance with 99.21 % detection rate, 99.07% precision, 99.15% recall, 99.18% F1 measure, and it provides an effective way of identifying threats like DDoS, SQL Injection, and Ransomware. In conclusion, Cyber-CapG provides a solid solution for traffic surveillance in IoT/IIoT, providing the high detection and prevention of various cyber threats, as well as optimizing the cybersecurity and reliability of industrial networks. © 2025 IEEE.

关键词： Industrial internet of things (IIoT)

来源：评论

学校读者我要写书评

暂无评论

Semantic Segmentation Based Real-Time Traffic Monitoring via Res-UNet Classifier and Kalman Filter

引用

SN computer Science 2025年第1期6卷 1-17页

作者： Qureshi, Asifa Mehmood Algarni, Asaad Aljuaid, Hanan Alatiyyah, Mohammed Hamad Alnowaiser, Khaled Jalal, Ahmad Department of Creative Technologies Air University E-9 Islamabad 44000 Pakistan Department of Computer Sciences Faculty of Computing and Information Technology Northern Border University Rafha 91911 Saudi Arabia Department of Computer Sciences College of Computer and Information Sciences Princess Nourah Bint Abdulrahman University P.O. Box 84428 Riyadh 11671 Saudi Arabia Department of Computer Science College of Sciences and Humanities in Aflaj Prince Sattam Bin Abdulaziz University Al-Kharj Saudi Arabia Department of Computer Engineering College of Computer Engineering and Sciences Prince Sattam Bin Abdulaziz University Al-Kharj 11942 Saudi Arabia Department of Computer Science Air University E-9 Islamabad 44000 Pakistan

In computer vision, accurate vehicle recognition and tracking is a challenging research topic. Manual surveillance systems are cumbersome, labor-intensive, and inefficient in today's congested traffic environment. Automated traffic monitoring systems are needed in today's world because they provide potential solutions to these issues. However, there are still some challenges in automating these systems which need to be addressed. One of the challenges is the extraction of foreground from a complex environment that only contains vehicles along with spatial and temporal information. In this paper, a semantic segmentation-based foreground extraction model is proposed to overcome these challenges. After segmentation, a blob detection technique is applied to detect each vehicle. To overcome the issue of vehicle-to-vehicle occlusion we extracted Harris corner points and computed the separation line based on the vehicle geometry coordinates. Furthermore, we introduce an ID assignment mechanism based on VGG-16 extracted features. Based on the detections, the number of cars is also counted for each frame. For tracking, we first applied Farneback optical flow algorithm to detect motion in vehicles. Only those vehicles were passed to the tracking algorithm in which motion was detected. To estimate each vehicle's location in succeeding image frames Kalman filter was implemented. A vehicle reidentification and ID recovery module are also proposed to enhance the tracking outcome. Vehicle trajectories based on vehicle location midpoint were also estimated. Extensive experiments are used to validate the proposed system and achieved a precision score of 0.965 and 0.931 over Roundabout and AU-AIR datasets for vehicle detection. However, for vehicle tracking the precision is 0.948 over the Roundabout dataset and 0.839 for the AU-AIR dataset. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Blob detection Detection and tracking Kalman filter Occlusion handling Res-Unet Semantic segmentation

来源：评论

学校读者我要写书评

暂无评论

MESC-3D:Mining Effective Semantic Cues for 3D Reconstruction from a Single Image

arXiv

引用

arXiv 2025年

作者： Li, Shaoming Cai, Qing Kong, Songqi Tan, Runqing Tong, Heng Qiu, Shiji Jiang, Yongguo Liu, Zhi Faculty of Computer Science and Technology Ocean University of China China School of Information Science and Engineering Shandong University China

Reconstructing 3D shapes from a single image plays an important role in computer vision. Many methods have been proposed and achieve impressive performance. However, existing methods mainly focus on extracting semantic information from images and then simply concatenating it with 3D point clouds without further exploring the concatenated semantics. As a result, these entangled semantic features significantly hinder the reconstruction performance. In this paper, we propose a novel single-image 3D reconstruction method called Mining Effective Semantic Cues for 3D Reconstruction from a Single Image (MESC-3D), which can actively mine effective semantic cues from entangled features. Specifically, we design an Effective Semantic Mining Module to establish connections between point clouds and image semantic attributes, enabling the point clouds to autonomously select the necessary information. Furthermore, to address the potential insufficiencies in semantic information from a single image, such as occlusions, inspired by the human ability to represent 3D objects using prior knowledge drawn from daily experiences, we introduce a 3D Semantic Prior Learning Module. This module incorporates semantic understanding of spatial structures, enabling the model to interpret and reconstruct 3D objects with greater accuracy and realism, closely mirroring human perception of complex 3D environments. Extensive evaluations show that our method achieves significant improvements in reconstruction quality and robustness compared to prior works. Additionally, further experiments validate the strong generalization capabilities and excels in zero-shot preformance on unseen classes. Code is available at https://***/QINGQINGLE/MESC-3D. Copyright © 2025, The Authors. All rights reserved.

关键词： 3D reconstruction

来源：评论

学校读者我要写书评

暂无评论

Towards distinguishing trust based attacks in an IoV network

引用

Journal of King Saud University - computer and information Sciences 2025年第4期37卷 1-15页

作者： Wang, Yingxun Mahmood, Adnan Mohd Sabri, Mohamad Faizrizwan Zen, Hushairi Faculty of Engineering Universiti Malaysia Sarawak Sarawak Kota Samarahan 94300 Malaysia Faculty of Computer and Information Engineering Qilu Institute of Technology Jinan 250200 China School of Computing Macquarie University Sydney 2109 NSW Australia Faculty of Engineering and Technology i-CATS University College Sarawak Kuching 93350 Malaysia

The paradigm of the Internet of Vehicles (IoV) promises significant innovative advancements and is receiving an increased attention of the researchers from both academia and industry. The IoV creates an intelligent network between different road entities and their corresponding roadside infrastructure via state-of-the-art sensing and communication technologies and, therefore, facilitates in addressing a range of safety-critical and non-safety vehicular applications. Accordingly, the security of an IoV network is of importance with internal attacks being of considerable concern. Although the conventional cryptography-based schemes are extremely intelligent in tackling the external attacks, trust has been recently employed to handle the internal attacks. Nevertheless, trust-based attacks also present a formidable challenge. Therefore, in this paper, an IoV-based trust management heuristic has been envisaged that takes into account both direct trust and indirect trust to ascertain the behaviors of the vehicles vis-à-vis time in a bid to detect various trust-based attacks, i.e., zig-zag attacks, self-promoting attacks, on-off attacks, and opportunistic attacks, along with the attackers’ multiple attacking strategies. The experimental findings further demonstrate that the envisaged IoV-based trust management heuristic exhibits prompt and accurate detection of the trust-based attacks in contrast to the state-of-the-art trust management mechanisms. © The Author(s) 2025.

关键词： Internet of vehicles Non-safety vehicular applications Safety-critical vehicular applications Trust-related attacks Vehicles’ behaviors

来源：评论

学校读者我要写书评

暂无评论

A Study of Predicting Social Networking Sites(SNS) Media Disorder Survey Data Using a Machine Learning Approach

A Study of Predicting Social Networking Sites(SNS) Media Dis...

引用

Multi-Agent Systems for Collaborative Intelligence (ICMSCI), International Conference on

作者： Jerin Akther Munia K.B.M. Tahmiduzzaman Izazul Haque Saad Md Mynul Islam Abdul Hady Akash Md. Thoufiq Zumma Department of Computer Science and Engineering Daffodil International University Dhaka Bangladesh Faculty of Engineering and Information Technology University of Technology Sydney Sydney Austrialia Department of Computer Science and Engineering University of Asia Pacific Dhaka Bangladesh

ISBN: (数字)9798331509828

ISBN: (纸本)9798331509835

This study uses survey data and machine learning algorithms to forecast social media disorder in people. A total of 600 individuals answered questions on their social media usage patterns, internet habits, demographics, and signs of social media disorder. This dataset was used to train a number of machine learning models, such as Gaussian Naive Bayes (GaussianNB), Random Forest, K-Nearest Neighbors (KNN), Decision Tree, and Support Vector Machine (SVM). According to the findings, the GaussianNB model had the greatest accuracy of 97%, followed by SVM (94%), Random Forest (91%), Decision Tree (92%), and KNN (90%). These results demonstrate how well machine learning can identify those who may be at risk for social media disorder. In order to improve prediction performance, future research will concentrate on increasing feature size, diversifying datasets, and refining models. Furthermore, incorporating these models into user-friendly apps and intervention programs may offer useful resources for the early diagnosis and treatment of social media illness.

关键词： Support vector machines Surveys Accuracy Machine learning algorithms Social networking (online) Nearest neighbor methods Predictive models Media Decision trees Random forests

来源：评论

学校读者我要写书评

暂无评论

Deep learning and ensemble methods for anomaly detection in ICS security

引用

International Journal of information technology (Singapore) 2025年第3期17卷 1761-1775页

作者： Hossain, Md. Alamgir Hasan, Tahmid Karovic, jr, Vincent Abdeljaber, Hikmat A. M. Haque, Md. Alimul Ahmad, Sultan Zafar, Aasim Nazeer, Jabeen Mishra, B.K. Department of Computer Science and Engineering State University of Bangladesh South Purbachal Kanchan Dhaka 1461 Bangladesh Department of Computer Science and Engineering Prime University Dhaka Bangladesh Department of Information Systems Faculty of Management Comenius Bratislava Odbojárov 10 Bratislava 25 82005 Slovakia Department of Computer Science Faculty of Information Technology Applied Science Private University Amman 11931 Jordan MEU Research Unit Middle East University Amman 11831 Lebanon Department of Computer Science Veer Kunwar Singh University Bihar Ara India Department of Computer Science College of Computer Engineering and Sciences Prince Sattam Bin Abdulaziz University Alkharj 11942 Saudi Arabia University Center for Research and Development (UCRD) Department of Computer Science and Engineering Chandigarh University Gharuan Punjab Mohali 140413 India Department of Computer Science Aligarh Muslim University U.P Aligarh 202002 India P.G. Department of Physics Veer Kunwar Singh University Bihar Ara India

This research addresses the escalating threats to industrial control systems by introducing a novel approach that combines deep learning for feature selection with a robust ensemble-based classification technique to enhance anomaly detection. Our method utilizes a tailored autoencoder architecture to efficiently select features, followed by a Random Forest classifier to ensure reliable and generalizable detection of anomalies. Evaluated on the HAI Security Dataset, the proposed approach demonstrates exceptional performance, with significant improvements in detecting anomalous activities in ICS. The results highlight the method’s potential to enhance ICS security by providing a scalable and adaptable framework that evolves with emerging threats. This study not only advances cybersecurity methodologies for industrial systems but also lays the groundwork for future research, emphasizing the need for innovative and effective security solutions in the face of evolving cyber challenges. © Bharati Vidyapeeth's Institute of computer Applications and Management 2024.

关键词： Anomaly detection in ICS Cyber-resilience in critical infrastructure Deep learning for feature selection Industrial control systems security Random forest classifier for ICS

来源：评论

学校读者我要写书评

暂无评论

A Reinforcement Learning-Based Image Mapping-Topology Optimization for Antenna Design

引用

IEEE Antennas and Wireless Propagation Letters 2025年

作者： Dou, Jiangling Gong, Hao Chen, Haokang Chen, Yujie Shen, Tao Song, Jian Kunming University of Science and Technology Faculty of Information Engineering and Automation Yunnan Key Laboratory of Computer Technologies Application Kunming650500 China Kunming University of Science and Technology Faculty of Information Engineering and Automation Kunming Yunnan650500 China College of Mechanical and Electrical Engineering Yunnan Electromechanical Vocational and Technical College China

A Reinforcement Learning (RL)-based Image Mapping Topology (IMT) optimization approach is proposed. To enable RL to optimize antenna topology and increase the efficiency of the optimization process. Two crucial components, Stratified Sampling Balance (SSB) and IMT are introduced. SSB is utilized to balance different types of data samples. IMT is utilized to facilitate the interaction between RL and simulation software, it also pretrains the actor network within the RL algorithm. This methodology allows the RL agent to acquire knowledge from the interaction of the antenna topology with the environment. Subsequently, the agent outputs antenna topology with the maximum reward signal. Compared with other machine learning-based optimization methods, the proposed method significantly alleviates the issue of learning inefficiency and reduces the exploration time for feasible solutions. To validate the effectiveness of the optimization method, a microstrip antenna is optimized. The measurement results show that the impedance bandwidth is doubled while maintaining the realized gain with fewer simulation samples and less optimization time. © 2025 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Research on Citrus Leaf Disease Recognition Using Class-Agnostic Contrastive Learning and Supervised Organizational Mapping

引用

IEEE Access 2025年

作者： Xiao, Wen Shang, Jinxia Li, Fan Ao, Ou Wang, Xin Tian, Shiyi Kunming University of Science and Technology Faculty of Information Engineering and Automation Kunming650500 China Kunming University of Science and Technology Yunnan Key Lab of Computer Technology Applications Kunming650500 China

Early detection of plant diseases is crucial for enhancing agricultural productivity and ensuring crop protection. While computer vision offers scalable alternatives to manual inspection, existing methods face two understudied challenges: (1) performance degradation in cluttered real-world conditions, and (2) ineffective feature separation for randomly distributed lesions with diverse symptoms. Notably, current research predominantly relies on controlled-environment datasets, lacking publicly accessible benchmarks with authentic field complexities. To address these, we present the first large-scale citrus leaf disease dataset (5,763 images, 6 categories) capturing complex field conditions including varying illumination, occlusion, multi-scale lesions, and deceptive backgrounds. Our key methodological innovations include a three-stage weakly supervised framework that systematically resolves background interference and feature redundancy: (i) Class-agnostic contrastive localization enabling precise lesion mapping without region annotations, (ii) Stable feature initialization through region-guided classification training, (iii) Sparse orthogonal mapping with experimentally-determined nodes introducing orthogonality constraints to enhance feature separation. Through feature alignment and feature fusion, an accuracy of 97.1% is achieved (+2.4% vs ResNet-18). Evaluations on citrus leaf and public datasets demonstrate superior accuracy and interpretable lesion localization, advancing plant disease diagnosis. © 2013 IEEE.

关键词： Citrus fruits

来源：评论

学校读者我要写书评

暂无评论

Understanding Multimodal LLMs Under Distribution Shifts: An information-Theoretic Approach

arXiv

引用

arXiv 2025年

作者： Oh, Changdae Fang, Zhen Im, Shawn Du, Xuefeng Li, Yixuan Department of Computer Sciences University of Wisconsin–Madison WI United States Faculty of Engineering & Information Technology University of Technology Sydney Sydney Australia

Multimodal large language models (MLLMs) have shown promising capabilities but struggle under distribution shifts, where evaluation data differ from instruction tuning distributions. Although previous works have provided empirical evaluations, we argue that establishing a formal framework that can characterize and quantify the risk of MLLMs is necessary to ensure the safe and reliable application of MLLMs in the real world. By taking an information-theoretic perspective, we propose the first theoretical framework that enables the quantification of the maximum risk of MLLMs under distribution shifts. Central to our framework is the introduction of Effective Mutual information (EMI), a principled metric that quantifies the relevance between input queries and model responses. We derive an upper bound for the EMI difference between in-distribution (ID) and out-of-distribution (OOD) data, connecting it to visual and textual distributional discrepancies. Extensive experiments on real benchmark datasets, spanning 61 shift scenarios empirically validate our theoretical insights. © 2025, CC BY.

关键词： information theory

来源：评论

学校读者我要写书评

暂无评论

SVIFNN: Robust Inpainting Fourier Neural Network for SST Scientific Visualization Image: Leveraging Significant Stability and Non-Significant Anomalies

引用

IEEE Transactions on Geoscience and Remote Sensing 2025年

作者： Zuo, Zijie Nie, Jie Wang, Xin Dong, Junyu Ocean University of China Faculty of Information Science and Engineering Qingdao266100 China Tsinghua University Department of Computer Science and Technology Beijing100084 China

The Scientific Visualization Images (SVI) of Sea Surface Temperature (SST) play a pivotal role as visual resources for investigating oceanographic processes. However, they are often plagued by extensive data gaps due to objective factors like cloud cover. Additionally, their content deviates significantly from ordinary images, posing challenges for conventional completion techniques. Given the intricate nature of marine systems, completing the visualization of sea surface temperature presents several challenges. Firstly, predicting missing segments relies not only on prominent patterns but also on subtle anomalies, which are often overlooked by methods focused on extracting prominent features. Secondly, these images exhibit chaos and lack clear semantics, making it difficult for methods primarily focused on semantic extraction to effectively complete them. To address these challenges, this study presents a novel method named the Inpainting Fourier Neural Network for SVI (SVIFNN). This approach employs a twin-stream architecture to highlight both significant stability and non-significant anomalies. Notably, it incorporates a "reverse attention mechanism" in the non-significant anomalies extraction stream to preserve unconventional information. Furthermore, by cascading Fourier neural operator (FNO), it leverages frequency domain characteristics to mitigate spatial chaos. Through a frequency domain feature extraction module, it achieves an adaptive fusion of significant stability and nonsignificant anomalies. Experiment results demonstrate SVIFNN’s superiority over State-Of-The-Art (SOTA) methods, particularly under a 68% missing rate condition. Significant improvements are observed in R2 (18.1%, 19.3%, and 21.8%) and reductions in RMSE (22.6%, 28.8%, and 23.8%) across different Noise-to-Signal (N/S) ratios of 0.1, 0.2, and 0.3, respectively, underscoring SVIFNN’s robustness in handling SVIs extensive data gaps. Adequate ablation experiments further validate the effec

关键词： Image enhancement

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：