检索结果-内蒙古大学图书馆

3rd International Conference on Intelligent Data Communication Technologies and Internet of Things, IDCIoT 2025

作者： Senthamizh, G. Anunya, G. Malathy, C. Jagadesh, R. Prithiv Sri Prasanth, C.S. Department of Networking and communication Engineering SRM Institute of Science and Technology Chennai India

ISBN: (数字)9798331527549

ISBN: (纸本)9798331527549

networking Intrusion Detection Systems (NIDS) are essential for identifying and mitigating cyberthreats in modern networking infrastructures. Modern NIDS still struggle with scalability, data imbalance, and high false positive rates despite significant advancements. This paper proposes a system that integrates state-of-the-art machine learning (ML) methods with Explainable AI (XAI) approaches to improve the efficacy, accuracy, and interpretability of NIDS. The study evaluates existing systems, identifies their flaws, and proposes an enhanced architecture that includes real-time threat adaptation, scalable processing, and an intuitive dashboard for security analysts. The experiment’s results demonstrate a significant drop in false positives, an increase in detection rates, and improved user insights into system selections. © 2025 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Disaster Assistance Intelligent Agent 14

Disaster Assistance Intelligent Agent

引用

14th International Conference on Computing Communication and networking Technologies, ICCCNT 2023

作者： Priyadarshi, Siddhant Rastogi, Mudit Gayathri, V.M. SRM Institute of Science and Technology Computer Science Engineering Chennai India SRM Institute of Science and Technology Department of Networking with Comunication Chennai India

ISBN: (纸本)9798350335095

Virtual characters are controlled by artificial intelligence in video games and movies. Controlling these characters when they are surrounded by people is known as crowd simulation. In this work, we propose to employ intelligent agents and decision trees to simulate a realistic scenario in which our agent will lead the human behavioural AI and identify the best route to the goal or the best circumstances to the objective avoiding all the obstacles in the path. This crowd simulation will be used to make intelligent agents which can find the best path during the situation of any disaster and show humans the safest path which the humans can follow to get to the nearest safe points. Multiagent simulation, commonly referred to as a Crowd and Agent Simulation, is an approach used to simulate many people, animals or other characters interacting with each other in one place. These actors are expected to pursue their objectives, engage in the environment and coordinate with each other. It is expected that the simulated actors will pursue their goals, interact with the environment, and communicate among themselves. Crowd simulations find utility in various areas, including the enhancement of training environments and virtual worlds, the optimization of architectural planning, and the refinement of AI-driven characters' driving behaviors in video games and movies. © 2023 IEEE.

关键词： Decision trees

来源：评论

学校读者我要写书评

暂无评论

AI-Driven Safety: Automated Helmet Detection and License Plate Recognition 9

AI-Driven Safety: Automated Helmet Detection and License Pla...

引用

9th International Conference on Communication and Electronics Systems, ICCES 2024

作者： Angayarkanni, S.A. Bathija, Nikhil Saketh Sourya, J. Shekhawat, Nikhil Singh Rajaram, V. Balakiruthiga, B. Srm Institute of Science and Technology Department of Networking and Communication Chennai India

ISBN: (纸本)9798350377972

This research study addresses the increasing need for automated traffic enforcement solutions to improve road safety, focusing on helmet detection and license plate recognition for motorcyclists. With manual enforcement limited by factors such as human error, fatigue, and the sheer volume of traffic, an AI-driven system offers a scalable alternative. Leveraging YOLOv8 for object detection and EasyOCR for text extraction, the proposed system automates the detection of helmet violations and vehicle identification through number plates. The model operates by first detecting helmet usage, and if absent, proceeds to recognize the vehicle's number plate. It performs reliably in various challenging conditions like poor lighting, occlusion, and motion blur, making it suitable for real-time traffic monitoring. The research also addresses ethical considerations around privacy and data protection, ensuring compliance with legal standards. By reducing manual intervention, this solution enhances traffic enforcement efficiency, promotes road safety, and provides a scalable approach to monitoring traffic violations. © 2024 IEEE.

关键词： Optical character recognition

来源：评论

学校读者我要写书评

暂无评论

DDoS Attack Detection System for IoT Enabled Smart City Applications with Correlation Analysis 4

DDoS Attack Detection System for IoT Enabled Smart City Appl...

引用

4th International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies, ICAECT 2024

作者： Nirosha, Veeramachaneni Hemamalini, V. Srm Institute of Science and Technology Department of Networking and Communication Kattankulathur India

ISBN: (纸本)9798350343670

The topic of cybersecurity remains a significant concern for industries involved in digital operations, as seen by the recurring increase in security incidents. The increasing use of Internet of Things (IoT) devices in various settings such as homes, offices, transportation, and healthcare has resulted in a corresponding rise in the frequency of malicious assaults. Machine Learning (ML) is frequently utilized for the purpose of detecting attacks on IoT devices. This study presents a novel approach to cybersecurity-focused network traffic correlation analysis and the detection of Distributed Denial of Service (DDoS) attacks using artificial intelligence techniques in the context of an IoT-enabled smart city, with the aim of promoting sustainability. A traffic analysis was conducted by utilizing a wireshark network analyser tool, which effectively improves data transmission. The experimental investigation was conducted to evaluate various performance metrics, including accuracy, precision, recall, attack detection rate and time. The proposed technique achieved anaccuracy rate of 96%, precision and recall rates are 93.6% and 91% respectively. © 2024 IEEE.

关键词： correlation analysis Cyber security DDoS attack smart city

来源：评论

学校读者我要写书评

暂无评论

Integrating Deep Q-Networks and YOLO with Autonomous Robots for Efficient Warehouse Management 5

Integrating Deep Q-Networks and YOLO with Autonomous Robots ...

引用

5th IEEE International Conference on Recent Trends in computer science and technology, ICRTCST 2024

作者： Vivek, N. Suriya, A. Kamala, V. Chinnasamy, N.V. Akshya, J. Sundarrajan, M. Choudhry, Mani Deepak SRM Institute of Science and Technology Department of Computational Intelligence Chennai India KGiSL Institute of Technology Department of Computer Science and Engineering Coimbatore India Karpagam Academy of Higher Education Department of Computer Science Coimbatore India SRM Institute of Science and Technology Department of Networking and Communications Chennai India KGiSL Institute of Technology Department of Information Technology Coimbatore India

ISBN: (纸本)9798350351378

Autonomous robotic cars' rapid, accurate, and efficient product handling and navigation are changing warehouse inventory management. These 24/7 smart vehicles save labor costs, human error, and operational inefficiencies. Time-consuming and incorrect manual scanning and recording is used for inventory management. Traditional autonomous car Q-learning algorithms may suffer in dynamic inventory circumstances and lack the complexity needed for real-time decision-making. Our research creates novel algorithms for autonomous robotic cars. The Deep Q-Network (DQN) we utilize for navigation adapts better to changing circumstances than ordinary Q-learning. YOLO object detection replaces slower QR code recognition for inventory monitoring. We use Decision Trees to find QR code data insights for decision-making. Our research results are promising. A greater cumulative reward curve over 100 simulated episodes gives DQN a 40% learning efficiency advantage over Q-learning. With 95% accuracy, YOLO outperformed Haar Cascade classifiers in QR code detection by 40% speed and 15% accuracy. Decision Trees identified QR code data 5.9% better than Rule-Based Systems, indicating inventory dependability. Autonomous robots can simplify inventory management in warehouses. Algorithmic technology allows these ambulances to drive intelligently, detect warehouse things faster and more accurately, and make decisions based on complicated data. These advanced artificial intelligence algorithms may make robotic systems more responsive and customized, changing inventory management subfields. These algorithms must be updated so autonomous cars can drive to meet logistical needs. © 2024 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

TextFormer: A Query-based End-to-end Text Spotter with Mixed Supervision

引用

Machine Intelligence Research 2024年第4期21卷 704-717页

作者： Yukun Zhai Xiaoqiang Zhang Xiameng Qin Sanyuan Zhao Xingping Dong Jianbing Shen School of Computer Science Beijing Institute of TechnologyBeijing 100081China Department of Computer Vision Technology Baidu Inc.Beijing 100193China

End-to-end text spotting is a vital computer vision task that aims to integrate scene text detection and recognition into a unified *** methods heavily rely on region-of-interest(Rol)operations to extract local features and complex post-processing steps to produce final *** address these limitations,we propose TextFormer,a query-based end-to-end text spotter with a transformer ***,using query embedding per text instance,TextFormer builds upon an image encoder and a text decoder to learn a joint semantic understanding for multitask *** allows for mutual training and optimization of classification,segmentation and recognition branches,resulting in deeper feature sharing without sacrificing flexibility or ***,we design an adaptive global aggregation(AGG)module to transfer global features into sequential features for reading arbitrarilyshaped texts,which overcomes the suboptimization problem of Rol ***,potential corpus information is utilized from weak annotations to full labels through mixed supervision,further improving text detection and end-to-end text spotting *** experiments on various bilingual(i.e.,English and Chinese)benchmarks demonstrate the superiority of our *** on the TDA-ReCTS dataset,TextFormer surpasses the state-of-the-art method in terms of 1-NED by 13.2%.

关键词： End-to-end text spotting arbitrarily-shaped texts transformer mixed supervision multitask modeling.

来源：评论

学校读者我要写书评

暂无评论

IoT and CNN Based Fire Detection And Prevention

IoT and CNN Based Fire Detection And Prevention

引用

2024 IEEE International Conference on Intelligent Computing and Emerging Communication Technologies, ICEC 2024

作者： Kavyasri, V.M. Krishnan Sudhakar, Nakul Malleswari, Tyj Naga Raghavendra, Sreesh Sundareshwaran, R.Kavin Venkat Phaneesh, Y. SRM Institute of Science and Technology Department of Networking and Communications India SRM Institute of Science and Technology Department of Computing Technologies India

ISBN: (纸本)9798331508432

Fire incidents pose significant risks to life and property, necessitating the development of advanced fire detection and suppression systems. This paper presents an innovative approach that combines a Convolutional Neural Network (CNN)- based image processing model with hardware integration for real- time fire detection and immediate response actions. The model is based on Yolo-v4 and techniques Cross validation and Grey Level Co-occurrence and median filtering are used for training of the model. The system utilizes a webcam for image capture, employing a CNN model trained on a large dataset to accurately identify fire conditions. Hardware components including temperature and humidity sensors (DHT11) and flame sensors complement the image processing for enhanced accuracy and reliability. The smoke sensors are usually used widely, but the flame sensors have a quicker response time. The proposed system demonstrates exceptional performance, achieving a detection accuracy 98%. Upon detecting fire, the system triggers an alarm through a buzzer for immediate evacuation alerts and activates a water motor for fire suppression. This integrated approach ensures rapid response to fire incidents, minimizing potential damages and enhancing overall safety. Key system features include scalability, adaptability, and user-friendliness, suitable for diverse environments like residential buildings, industries, and public spaces. Future upgrades might integrate advanced cameras, use machine learning for adaptive responses, and implement continuous monitoring for false positives, enhancing system capabilities. Overall, the developed system represents a significant advancement in leveraging technology for fire safety, with promising implications for mitigating fire-related risks effectively. © 2024 IEEE.

关键词： Median filters

来源：评论

学校读者我要写书评

暂无评论

URL Shield: Protecting Users from Phishing Attacks using Flask and ML 3

URL Shield: Protecting Users from Phishing Attacks using Fla...

引用

3rd International Conference for Innovation in technology, INOCON 2024

作者： Sandhya Godwin Ponsam, J. Srm Institute of Science and Technology Department of Networking and Communications Chennai India

ISBN: (纸本)9798350381931

Cybersecurity threats, especially phishing attacks, continue to pose a significant risk to internet users. In response to this growing menace, we present "PhishShield,"a robust and scalable solution that leverages machine learning within the Flask framework to detect phishing URLs effectively. Phish Shield incorporates a comprehensive set of features, including IP address analysis, URL characteristics, and domain-related attributes, which serve as the foundation for our machine learning models. Specifically, we employ Bernoulli naive bayes (Bernoulli NB) and Multinomial naive bayes (Multinomial NB) algorithms to classify URLs into three categories: Phishing, Legitimate, or Suspicious. Our system follows a systematic approach: it first collects and preprocesses datasets, and subsequently trains and tests the machine learning models. Upon successful training, PhishShield generates a pickle file, which can be used to determine whether a URL is susceptible to a phishing attack. The Flask-based web interface offers a user-friendly platform for users to input URLs and obtain instant assessments. Phish Shield's utilization of static CSS and Bootstrap for web design ensures an intuitive and visually appealing user experience. The templates "base,""check,""home,""index,"and "result"enhance the usability and interactivity of the system. PhishShield, we aim to empower users to navigate the web securely, providing a critical defense against phishing attacks. Our project not only showcases the capabilities of Flask and machine learning but also contributes to the ongoing battle against cyber threats. © 2024 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Edge Based Smart Energy Management System Using Object Detection

Edge Based Smart Energy Management System Using Object Detec...

引用

International Conference on Advances and Applications of Artificial Intelligence and Machine Learning, ICAAAIML 2023

作者： Bennish, M. Saravanan, M. Department of Networking and Communications SRM Institute of Science and Technology Chennai Kattankulathur India

ISBN: (纸本)9789819795062

With rising demand for energy efficiency and sustainability, advanced technologies to optimize energy consumption in various domains are in high demand. This paper proposes an Edge-based Smart Energy Management System that intelligently monitors and controls energy usage in real-time using object detection techniques. This system employs a distributed architecture to efficiently manage energy consumption by combining the power of edge computing and computer vision. Smart computer vision devices are deployed at the edge to collect real-time data from the environment. These devices use YOLOv8 object detection algorithms to identify and track energy-intensive objects such as appliances like lights, and fans. The detected objects information is then processed and analyzed at the edge devices, which generate the corresponding threshold value for managing the electrical appliances. These processed data will be sent to the cloud infrastructure for further analysis and decision-making. Through the integration of object detection and YOLOv8 machine learning techniques, this system offers several key benefits. Firstly, it enables real-time monitoring of energy consumption at a granular level, allowing users to identify the number of persons and take appropriate actions to reduce electricity consumption. Secondly, the system provides accurate energy forecasting, enabling proactive load management and efficient allocation of energy resources. The proposed system will implement the computer vision in real-time monitoring and evaluation to assess its performance, scalability, and energy-saving potential. The evaluation results will demonstrate the proposed system’s effectiveness in optimizing energy usage, reducing costs, and contributing to a more sustainable environment. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Email Phishing Detection Using AI and ML 2nd

Email Phishing Detection Using AI and ML

引用

2nd International Conference on Deep sciences for Computing and Communications, IconDeepCom 2023

作者： Sarkar, Sohan Yadav, Ankit Balachander, T. Department of Networking and Communications SRM Institute of Science and Technology Chennai India

ISBN: (纸本)9783031689048

Phishing is the act of attackers sending malicious emails to receivers in an effort to trick them into falling for a con. Normally, the intention is to persuade users to provide private information, such as system logins or financial data. Our project investigates such email phishing attacks using AI and ML to get a reasonable conclusion about which algorithm can spot these attacks the most successfully. In terms of cybersecurity, a phishing attack is a cybercrime that aims to obtain a user’s personal information in order to carry out some destructive actions. This attack’s effects could lead to account takeover, privilege escalation, and other issues. And in order to lessen it, this study provides information on how to spot phishing emails so that businesses can correctly deal with them. Natural Language Processing (NLP) [26], logistic regression, and fundamental AI ideas like CNN [27] are used, along with machine learning algorithms like KNN, Naive Bayers, and these. This study aims to identify the optimal machine learning algorithm that would provide the highest level of accuracy when it comes to phishing email detection. "To improve the accuracy of detecting phishing emails we have implemented CNN which is an effective approach to fulfill the target. CNN [27] can learn to recognize tiny patterns and traits that may be challenging for humans to notice, they can be a great tool for detecting phishing emails. Handling this technique is very important, hence LSTM is introduced". The CNN model’s LSTM [27] can be thought of as its brain. LSTM (Long Short-Term Memory) [27] is a form of recurrent neural network (RNN) that is frequently employed in tasks involving sequence prediction and natural language processing (NLP) [26]. It is intended to solve the vanishing gradients issue, which can arise in conventional RNNs when the gradients are extremely small and make the network struggle to learn long-term dependencies. By understanding and executing everything we have obse

关键词： Phishing

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：