检索结果-内蒙古大学图书馆

Hybrid CNN-ViT architecture to exploit spatio-temporal feature for fire recognition trained through transfer learning

引用

Multimedia Tools and Applications 2025年第8期84卷 4703-4732页

作者： Shahid, Mohammad Wang, Hong-Cyuan Chen, Yung-Yao Hua, Kai-Lung Dept. of Computer Science and Information Engineering National Taiwan University of Science and Technology Taipei Taiwan Dept. of Electronic and Computer Engineering National Taiwan University of Science and Technology Taipei Taiwan

Fires are becoming one of the major natural hazards that threaten the ecology, economy, human life and even more worldwide. Therefore, early fire detection systems are crucial to prevent fires from spreading out of control and causing destruction. Based on vision sensors, many fire detection techniques have evolved with the recent surge of curiosity in deep learning, which exploits the spatial features of individual images. However, fire can take different forms, scales, and combustion materials can produce different colors, making accurate fire detection from an image challenging. Small fires captured from long-distance cameras lack salient features, further complicating detection. This paper proposes a hybrid structure that uses attention-enhanced convolutional neural networks and vision transformers (CNN-ViT) to detect fire. The proposed CNN-ViT first pays spatial attention to every frame and then aggregates temporal contextual information from neighboring frames to improve detection performance. Due to the limited availability of training fire datasets, the study employs deep transfer learning for feature extraction using pre-trained CNN. We used various metrics to examine the efficacy of the proposed approach. The results showed that the CNN-ViT method outperformed previous models based on spatial-temporal characteristics by achieving a relative improvement in accuracy and F1 score. The satisfactory results on images contaminated with different intensities of noise confirm the robustness of the approach. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Fires

来源：评论

学校读者我要写书评

暂无评论

MOPO-HBT: A movie poster dataset for title extraction and recognition

引用

Multimedia Tools and Applications 2024年第18期83卷 54545-54568页

作者： Ghosh, Mridul Roy, Sayan Saha Banik, Bivan Mukherjee, Himadri Obaidullah, Sk Md Roy, Kaushik Dept. of Computer Science Shyampur Siddheswari Mahavidyalaya Howrah India Dept. of Computer Science & Engineering Aliah University Kolkata India Dept. of Electronics & Electrical Communication Engineering IIT Kharagpur Kharagpur India TISA Lab Dept. of Computer Science West Bengal State University Barasat India

Real-world images often encompass embedded texts that adhere to disparate disciplines like business, education, and amusement, to name a few. Such images are graphically rich in terms of font attributes, color distribution, foreground-background similarity, and component organization. This aggravates the difficulty of recognizing texts from these images. Such characteristics are very prominent in the case of movie posters. One of the first pieces of information on movie posters is the title. Automatic recognition of movie titles from images can aid in efficient indexing as well as information conveyance. However, it is accompanied by other texts like the names of actors, producers, taglines, dates, etc. Though the organization of components is somewhat similar across different film industries like Tollywood (West Bengal), Bollywood (Mumbai), and Hollywood (Los Angeles), the graffiti patterns differ in multifarious instances. To address the problem of movie title understanding, we propose a dataset named MOvie POsters-Hollywood Bollywood Tollywood (MOPO-HBT) that encompasses movie posters from the aforementioned industries. The entire dataset is publicly available (http://***/11564) for research purposes. The baseline results of title identification and recognition were obtained with a CNN-based (Convolutional Neural Network) approach, wherein the titles were extracted using the M-EAST (Modified-Efficient and Accurate Scene Text) detector model. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2023.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

Moth flame optimization algorithm for efficient feature selection and hybrid classification technique for intrusion detection system

引用

Multimedia Tools and Applications 2025年 1-21页

作者： Keerthanarani, U. Selvakumar, V. Singaravelan, S. Raja, S. Edwin Thenmozhi, K. Shunmugam, D. Arun Dept. of Computer Science and Engineering P.S.R Engineering College Sivakasi India Dept.Of Computer Science and Engineering Vel Tech Rangarajan Dr Sagunthala R & d Institute of Science and Technology Chennai Avadi India Dept. Of Computer Science and Engineering Coimbatore Institute of Engineering and Technology Coimbatore India Dept. Of Electronics and Communication Engineering P.S.R Engineering College Sivakasi India

The rapid expansion of the Internet of Things (IoT) brings numerous benefits but further presents fresh difficulties, especially in terms of security. The distributed and interconnected nature of IoT devices makes them susceptible to various security threats. An Intrusion Detection System (IDS) plays a crucial role in safeguarding IoT networks by monitoring, detecting, and responding to potential security incidents. The Internet of Things (IoT) connects billions of devices, sensors, and systems, transforming industries and everyday life. However, this extensive interconnectivity also introduces significant security risks, making Intrusion Detection systems (IDS) a critical component for safeguarding IoT environments. Feature selection is a crucial aspect of building effective Intrusion Detection systems (IDS). Selecting relevant features not only helps in achieving better classification accuracy but also contributes to reducing false alarm rates. In the proposed system a moth flame optimizer is utilized to select the efficient features from the dataset, overall accuracy of over 97%. Further it is classified employing Random Forest and J48 classifier which attains high classification accuracy and low false alarm rate. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

AI in CS Education: Opportunities, Challenges, and Pitfalls to Avoid

引用

ACM Inroads 2024年第3期15卷 52-57页

作者： Vahid, Frank Dept. of Computer Science and Engineering University of California Riverside United States

Csinstructors realize education may be impacted by generative AI (artificial intelligence). This article describes (1) opportunities, like 24/7 help or auto-grading, (2) challenges, like increased cheating or student isolation, and (3) pitfalls to avoid, like inequitable access or false cheating accusations. The descriptions come from 15 CS educators and practitioners, from diverse institutions, who met at a 2023 NSF workshop on AI in CS education. The article aims to assist the CS community to take advantage of generative AI to improve courses, embark on new research, treat students equitably, and ultimately improve CS education for instructors and students. © 2024 Association for Computing Machinery. All rights reserved.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Enacting Rural Empowerment: Implementing Cooperative Commerce in Communities using Machine Learning 2

Enacting Rural Empowerment: Implementing Cooperative Commerc...

引用

2nd IEEE International Conference on Networking and Communications, ICNWC 2024

作者： Skandarsini, R.R. Sangavi, M.S. Harika, Ravitla Varun, C.M. RMK Engineering College Dept. of Computer Science and Business Systems Thiruvallur India

ISBN: (纸本)9798350365269

In order to revitalize rural economies through the empowerment of self-help organizations and the promotion of economic growth, this study offers a revolutionary cooperative commerce platform. The platform integrates cutting-edge features such as digital ledger and inventory management for sellers, demand forecasting, and dynamic pricing, thereby fostering optimized production and efficient marketplaces. Central to the platform's efficacy is the implementation of user segmentation, which categorizes users based on income and spending behaviors, enabling targeted marketing strategies and continuous application improvements to enhance user engagement and profitability. Leveraging machine learning algorithms, personalized user experiences are cultivated, elevating buyer satisfaction and overall engagement. Additionally, administrative tools provide valuable insights for informed decision-making. Preliminary results demonstrate improved financial management for self-help groups, heightened economic activities, and an enriched user-centric experience. Identified challenges and proposed future enhancements underscore the platform's commitment to adaptability and responsiveness to the evolving needs of rural communities, ensuring sustained relevance and impact. © 2024 IEEE.

关键词： Predictive analytics

来源：评论

学校读者我要写书评

暂无评论

Vyoma-ADS : An AI-Driven Tool for Adaptive Monitoring and Anomaly Detection for Space Agriculture

Vyoma-ADS : An AI-Driven Tool for Adaptive Monitoring and An...

引用

2024 IEEE International Conference on Communication, Computing and Signal Processing, IICCCS 2024

作者： Das, Rajdeep Sharma, Vaishnavi Kumari, Priyanka Raj, Abhishek Das, Sudipta De, Indrajit Institute of Engineering and Management Dept. of Computer Science and Business Systems Kolkata India

ISBN: (纸本)9798350390759

As humanity ventures further into space exploration, securing sustainable resources becomes paramount. The success of space farming depends not only on the ability to cultivate crops but also on the effective management of environmental conditions. Introducing Vyoma-ADS: an innovative system that integrates IoT technology with an adaptive Anomaly Detection System, specifically designed for space agriculture. Our study underscores the crucial role of environmental monitoring within space shuttles, where even minor fluctuations can have a significant impact on plant growth and productivity. To tackle these challenges, we designed a compact IoT-enabled system to monitor the moisture content in the soil, the temperature and humidity of the surrounding environment, and the ppm value of unwanted gases in the atmosphere. Through rigorous testing and analysis, we established optimal thresholds conducive to plant growth in the unique conditions of space. Recognizing the inherent variability and occasional anomalies in environmental data, we investigated three distinct machine learning models for anomaly detection: Isolation Forest, One-Class Support Vector Machines (OCSVM) and Density-Based Spatial Clustering of Applications with Noise (DBSCAN). Our findings reveal that DBSCAN outperforms other techniques in identifying anomalies amidst fluctuating data. Additionally, we detail our approach to data visualization using a SaaS cloud platform, which enhances the comprehensibility of the study and facilitates better decision-making processes. This paper demonstrates how the integration of IoT monitoring and advanced anomaly detection methods can significantly improve the reliability and efficiency of space agriculture. By leveraging these technologies, we aim to create a robust system capable of maintaining optimal growing conditions for plants in space. © 2024 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Content Based Classification of Short Messages using Recurrent Neural Networks in NLP

Content Based Classification of Short Messages using Recurre...

引用

2024 International Conference on Artificial Intelligence, computer, Data sciences, and Applications, ACDSA 2024

作者： Bennet, Deepthi Tabitha Bennet, Preethi Samantha Thiagarajan, Priya Sundarakantham, K. Thiagarajar College of Engineering Dept. of Applied Mathematics and Computational Science Madurai India Thiagarajar College of Engineering Dept. of Computer Science and Business Systems Madurai India Thiagarajar College of Engineering Dept. of Computer Science and Engineering Madurai India

ISBN: (纸本)9798350394528

Short message spam poses a significant threat for all mobile phone users, as it can act as an efficient tool for cyberattacks including spreading malware and phishing. Traditional anti-spam measures are only minimally effective against modern spammers. Intelligent analysis of the content to categorise the messages as Spam (unwanted and unsolicited) or Ham (useful messages) is therefore essential to safeguard the user from such attacks. Various artificial intelligence (AI) techniques are proving to be useful in the analysis of the content of such short messages to classify and filter spam. We have trained, validated and tested seven such AI techniques on the SMS spam collection dataset to identify the best model for designing and developing a content based classification system. Recurrent Neural Networks (RNN) have shown the highest performance metrics (Test Accuracy: 99.28%) and hence our proposed system includes a RNN model for classification. A web app of this system has also been deployed where a single SMS can be input and the designed system can classify it as Spam or Ham. The designed system is compared against existing systems and is found to be significantly better. © 2024 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Artificial Intelligence and Machine Learning Algorithms to Analyse Integrated Nutrient Management Strategies for Optimal Yield and Quality of Oilseeds Crops for Sustainable Development 4

Artificial Intelligence and Machine Learning Algorithms to A...

引用

4th International Conference on Soft Computing for Security Applications, ICSCSA 2024

作者： Mithra, C. Dar, Showkat A. Preethi, A. Dept. of Computer Science and Engineering Chennai India Githam University Dept. of Computer Science and Engineeing India Dept. of Computer Science and Engineering Chidambaram India

ISBN: (纸本)9798331515720

The main objective of this research is to deduce the efficacy of integrated nutrient management (INM) technologies in production of oilseed crops for sustainable development. A great amount of experience is needed in selecting the most effective INM strategy. A new recommendation system to circumnavigate this issue is proposed. It lets farmers decide on the best INM strategy to maximize oilseed crop yield and quality. This system is built on the techniques of advanced machine Learning (ML) and aritifical Intelligence (AI). Oilseed crop date in Tamil Nadu from 1961 to 2019 was used to develop the proposed algorithm. The proposed algorithm for crop yield prediction (CYP) which includes a Soft Voting Ensemble Classifier with weights (SVECWW), a Soft Voting Ensemble Classifier without weights (SVECWOW) along with the SVM technique are compared and contrasted with existing algorithms and also proves that SVECWW outperforms other ML algorithms with an accuracy rate of 97.2%. Furthermore, the Stacked Generalization Ensemble model is employed and compared with another Deep Neural Network (DNN) for the INM crop recommendation system which offers a simple graphical user interface (GUI) for farmers to use and received an accuracy of 97.5%. This GUI enables farmers to access valuable information such as the optimal timing for cultivating oilseed crops, the appropriate types and quantities of organic manures, inorganic fertilizers, and bio-fertilizers required for successful oilseed crop production. The study shows, on its whole, how to create tailored recommendation systems for farmers using GUI models with artificial intelligence and machine learning algorithms. Implementing these systems is expected to significantly improve oilseed crop production and quality significantly, benefiting the whole agricultural sector for sustainable development. Artificial intelligence (AI) makes a recommendation system more accurate and adaptable by looking through complex datasets and patterns

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

J-NECORA: A Framework for Optimal Resource Allocation in Cloud-Edge-Things Continuum for Industrial Applications With Mobile Nodes

引用

IEEE Internet of Things Journal 2025年第11期12卷 16525-16542页

作者： Pettorali, Marco Righetti, Francesca Vallati, Carlo Das, Sajal K. Anastasi, Giuseppe Univ. of Pisa Dept. of Information Engineering Italy Missouri Univ. of Science and Technology Dept. of Computer Science RollaMO United States

In the Industrial Internet of Things (IIoT) landscape, where the Cloud-to-Things Continuum (C2TC) paradigm is now a reality, industrial applications need to cope with highly heterogeneous network and computing resources. Moreover, many industrial applications also involve Mobile Nodes (MNs). Efficient allocation of network and computing resources to meet the stringent requirements of such applications is often a very challenging task. In this paper, we propose J-NECORA (Joint NEtwork and COmputing Resource Allocation), a comprehensive analytical framework to derive the optimal joint allocation of network and computing resources in the C2TC, that guarantees the application requirements, even in the presence of MNs. If the optimal allocation does not exist, J-NECORA provides a best-effort allocation that minimizes the delay. The proposed framework is highly modular and flexible, and can be easily customized to different scenarios. We validated its effectiveness through simulation experiments, focusing on a specific use case. For the considered scenarios, the performance predicted by our framework matches quite closely the simulation results. © 2014 IEEE.

关键词： Cloud platforms

来源：评论

学校读者我要写书评

暂无评论

Bidirectional Multi-Step Domain Generalization for Visible-Infrared Person Re-Identification

Bidirectional Multi-Step Domain Generalization for Visible-I...

引用

2025 IEEE/CVF Winter Conference on Applications of computer Vision, WACV 2025

作者： Alehdaghi, Mahdi Shamsolmoali, Pourya Cruz, Rafael M. O. Granger, Eric Livia Ills Dept. of Systems Engineering Ets Montreal Canada University of York Dept. of Computer Science United Kingdom

ISBN: (纸本)9798331510831

A key challenge in visible-infrared person re-identification (V-I ReID) is training a backbone model capable of effectively addressing the significant discrepancies across modalities. State-of-the-art methods that generate a single intermediate bridging domain are often less effective, as this generated domain may not adequately capture sufficient common discriminant information. This paper introduces Bidirectional Multi-step Domain Generalization (BMDG), a novel approach for unifying feature representations across diverse modalities. BMDG creates multiple virtual intermediate domains by learning and aligning body part features extracted from both I and V modalities. In particular, our method aims to minimize the cross-modal gap in two steps. First, BMDG aligns modalities in the feature space by learning shared and modality-invariant body part prototypes from V and I images. Then, it generalizes the feature representation by applying bidirectional multi-step learning, which progressively refines feature representations in each step and incorporates more prototypes from both modalities. Based on these prototypes, multiple bridging steps enhance the feature representation. Experiments11Our code is available at: ***/BMDG conducted on V-I ReID datasets indicate that our BMDG approach can outperform state-of-the-art part-based and intermediate generation methods, and can be integrated into other part-based methods to enhance their V-I ReID performance. © 2025 IEEE.

关键词： Multi-task learning

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：