检索结果-内蒙古大学图书馆

2nd International Conference on Data Analytics, Computing and Artificial Intelligence, ICDACAI 2023

作者： Jiang, Tianyang Zhu, Yuke Feng, Haojun University of Shanghai for Science and Technology Usst Dept. School of Optical-Electrical and Computer Engineering Shanghai China Zhejiang Gongshang University Dept. School of Information and Electronic Engineering Hangzhou China Jurong Country Garden School Jurong China

ISBN: (纸本)9798350339765

Sounds carry a huge amount of information of our daily life and physical events. We hear sounds in complex environment and from increased amounts of electric devices, so audio recognition is definitely an important technology in human-computer interaction. In recent years, speech recognition technology developed rapidly. Summarizing recognition methods is necessary when there are too many methods to choose. This paper introduces some of the advanced technologies, Recurrent Neural Network, Convolutional Neural Network, Linear Regression, Multilayer Perceptron, Random Forest and their principle. © 2023 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Copy-Move Forgery Detection in Digital Forensic Images Using CNN 7

Copy-Move Forgery Detection in Digital Forensic Images Using...

引用

7th International Conference on computer science and engineering, UBMK 2022

作者： Kaya, Mustafa Sani, Khalid Jibril Karakus, Serkan Firat University Dept. of Digital Forensics Engineering Elazig Turkey Federal University of Kashere Dept. of Computer Science Gombe Nigeria

ISBN: (数字)9781665470100

ISBN: (纸本)9781665470100

Digital images are commonly used for sharing visual information and can be manipulated easily. The detection forgery in digital images has become a hot domain of research in digital image forensics due to the prevalent use of image editing tools for manipulating an image to conceal or distort information in the image. One of the most common image forgeries performed is the copy-move forgery. This type of forgery involves copying a segment of the image, which is then pasted to a different segment of the same image. The need for detecting whether an image is authentic becomes essential. The existing methods implemented for detecting image forgeries were based on traditional feature extraction algorithms such as block-based and key point-based algorithms. These traditional techniques employed produce a low-performance result. Deep learning techniques have proven to provide better performance in image processing tasks. In this research, a convolutional neural network based on a pre-trained ResNet50 network was proposed to detect copy-move forgeries in digital images. The proposed model uses the CoMoFoD image dataset in experimenting. The metric evaluation results achieved in the proposed model show that deep learning methods performance is more effective in digital image copy-move forgery detection. © 2022 IEEE.

关键词： Digital forensics

来源：评论

学校读者我要写书评

暂无评论

FireSense: AI-Based Wildfire and Smoke Detection in Forests

FireSense: AI-Based Wildfire and Smoke Detection in Forests

引用

2025 International Conference on New Trends in Computing sciences, ICTCS 2025

作者： Mughal, Abdul Basit Haider, Syed Mujtaba Khan, Rafi Ullah Ali Shah, Syed Mohsin Arain, Javiriya Hameed Department of Computer Science Bahria University Islamabad Pakistan Department of Mathmatics University of Pavia Pavia Italy Dept. of Electrical & Computer Engineering PAF-IAST Haripur Pakistan Faculty of Engineering University of Deusto Bilbao Spain Department of Computer Science COMSATS University Islamabad Pakistan

ISBN: (纸本)9798331523411

Forests are an essential part of the environment which plays a vital role in sustaining ecological balance and providing habitat to different wildlife. Wildfires are one of the serious threats to forests which create a devastating impact on the natural ecosystem. Early identification of smoke and wildfires will help to reduce damage and support a quick reaction. Earlier, wildfire and smoke detection in forests were very difficult due to the challenging forest terrain, low light conditions, and difficult weather *** the development in technology, several computer vision methods are combined with satellite and Unmanned Aerial Vehicle (UAV) data to correctly identify wildfires, but accuracy is still a concern with these models. This paper introduced a novel method to detect wildfire and smoke in forests utilizing fine-tuned YOLOv11n using an aerial vehicle dataset. The mAP of the fine-tuned YOLOv11n is 88.6%, with a precision of 100% and an F1-score of 85%. The number of parameters has decreased by 30% compared to its previous models, and the model size is only 5.5MB. ©2025 IEEE.

关键词： Object Detection smoke detection Unmanned Aerial Vehicle (UAV) Wildfire detection

来源：评论

学校读者我要写书评

暂无评论

Optimizing Multi-Class Classification in Educational Data with Ensemble Learning and Data Balancing Techniques

Optimizing Multi-Class Classification in Educational Data wi...

引用

Intelligent Data science Technologies and Applications (IDSTA), International Conference on

作者： Mohammad F. Al-Hammouri Ziad Akram Ali Hammouri Islam T. Almalkawi Ansam Lafee Dept. of Computer Engineering The Hashemite University Zarqa Jordan Dept. of Computer Science Middle East University Amman Jordan

ISBN: (数字)9798350354751

ISBN: (纸本)9798350354768

The field of education is increasingly embracing AI tools to improve student outcomes. This work aims to reduce academic failure in higher education by employing machine learning techniques to identify at-risk students early in their educational journey, enabling the implementation of supportive strategies to assist them. This study examines a dataset from a higher education institution and utilizes it to develop a classification model for predicting students’ academic performance. The problem is formulated as a multi-class classification task with three categories: Graduate, Enrolled, and Dropout, with a significant imbalance skewed toward the Graduate. To improve prediction accuracy toward the minority class, the data balancing technique SMOTE with Edited Nearest Neighbor (SMOTE-ENN) is applied. Three popular classification models—Random Forest, XGBOOST and CatBoost—are employed. The findings show that SMOTE-ENN significantly improves classification results. Moreover, XGBOOST demonstrated the highest accuracy (94.6%) in correctly identifying all classes, as evidenced by the confusion matrix evaluation, achieving the highest results compared to previous work in the literature. Implementing these models allows for accurate predictions of students’ performance and helps reduce dropout rates.

关键词： Performance evaluation Accuracy Machine learning algorithms Education Forestry Predictive models Data science Prediction algorithms Data models Ensemble learning

来源：评论

学校读者我要写书评

暂无评论

Deep Learning: A New Way of Looking into Fashion 4

Deep Learning: A New Way of Looking into Fashion

引用

4th International Conference on Advances in Computing, Communication Control and Networking, ICAC3N 2022

作者： Thomas, Josh Hanley Roopashree, S. Joy, Jacob Charles Abdullah Kollarath, Haseeb Siby, Arun Dept. CSE-AI and Ml Bengaluru India Rv Institute of Technology and Management Dept. Computer Science and Engineering Bengaluru India

ISBN: (纸本)9781665474368

Currently, Fashion-Fit using AI focuses on finding the perfect fit for users using GAN's (General Adversarial Net¬works) and various methods proposed in the Artificial Intelligence Industry. However, due to changes in fashion trends and various preferences of the "type of fit"users want for their clothing the prior methodology is not meant for the perfect fit. These days a lot of people wear tighter or even get an oversized fit based on the user's preference. In this paper we use Deep Learning techniques like CNNs and GANs to implement neural body fit by looking at various human body poses from large datasets and concatenating this data to make an accurate "desired"fit based on different human poses. This will set a new benchmark for virtual try-on. In the model being proposed, we are building Convolutional Neural Networks (CNNs) for body segmentation to make it easier to map a human pose. We will then map the human body using CNNs but not necessarily to find the "perfect fit". The CNNs will be used to make measurements that are similar to sizes available within the fashion industry allowing users to see how they would fit within those garments virtually. We will then use 2D renders, using uploaded pictures of the user from 4-5 different angles in order to allow the user to see how he/she would look in the outfit of their choice. These methods can give an accurate prediction helping every fashion e-commerce business. Personalized size and fit recommendations are critical for a fashion e-commerce business. Predicting the optimal fit increases consumer happiness while lowering expenses associated with size-related returns. © 2022 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

PhishGuard: A Multi-Layered Ensemble Model for Optimal Phishing Website Detection

arXiv

引用

arXiv 2024年

作者： Ovi, Md Sultanul Islam Rahman, Md. Hasibur Hossain, Mohammad Arif Dept. of Computer Science George Mason University FairfaxVA United States Dept. of Computer Science and Engineering Green University of Bangladesh Bangladesh

Phishing attacks are a growing cybersecurity threat, leveraging deceptive techniques to steal sensitive information through malicious websites. To combat these attacks, this paper introduces PhishGuard, an optimal custom ensemble model designed to improve phishing site detection. The model combines multiple machine learning classifiers, including Random Forest, Gradient Boosting, CatBoost, and XGBoost, to enhance detection accuracy. Through advanced feature selection methods such as SelectKBest and RFECV, and optimizations like hyperparameter tuning and data balancing, the model was trained and evaluated on four publicly available datasets. PhishGuard outperformed state-of-the-art models, achieving a detection accuracy of 99.05% on one of the datasets, with similarly high results across other datasets. This research demonstrates that optimization methods in conjunction with ensemble learning greatly improve phishing detection performance. Copyright © 2024, The Authors. All rights reserved.

关键词： Phishing

来源：评论

学校读者我要写书评

暂无评论

Context-Aware Neural Video Compression on Solar Dynamics Observatory 22

Context-Aware Neural Video Compression on Solar Dynamics Obs...

引用

22nd IEEE International Conference on Machine Learning and Applications, ICMLA 2023

作者： Khoshkhahtinat, Atefeh Zafari, Ali Mehta, Piyush M. Nasrabadi, Nasser M. Thompson, Barbara J. Kirk, Michael S. F. Da Silva, Daniel West Virginia University Dept. of Computer Science & Electrical Engineering WV United States West Virginia University Dept. of Mechanical & Aerospace Engineering WV United States Nasa Goddard Space Flight Center MD United States

ISBN: (纸本)9798350345346

NASA's Solar Dynamics Observatory (SDO) mission collects large data volumes of the Sun's daily activity. Data compression is crucial for space missions to reduce data storage and video bandwidth requirements by eliminating redundancies in the data. In this paper, we present a novel neural Transformer-based video compression approach specifically designed for the SDO images. Our primary objective is to efficiently exploit the temporal and spatial redundancies inherent in solar images to obtain a high compression ratio. Our proposed architecture benefits from a novel Transformer block called Fused Local-aware Window (FLa Win), which incorporates window-based self-attention modules and an efficient fused local-aware feed-forward (FLaFF) network. This architectural design allows us to simultaneously capture short-range and long-range information while facilitating the extraction of rich and diverse contextual representations. Moreover, this design choice results in reduced computational complexity. Experimental results demonstrate the significant contribution of the FLaWin Transformer block to the compression performance, outperforming conventional hand-engineered video codecs such as H.264 and H.265 in terms of rate-distortion trade-off. © 2023 IEEE.

关键词： FLaWin Neural Video Compression Solar Dynamics Observatory Swin Transformer

来源：评论

学校读者我要写书评

暂无评论

Batch Active Learning for Text Classification and Sentiment Analysis 22

Batch Active Learning for Text Classification and Sentiment ...

引用

3rd International Conference on Control, Robotics and Intelligent System, CCRIS 2022

作者： Sahan, Marko Smidl, Vaclav Marik, Radek Dept. of Computer Science Fee Ctu Prague Czech Republic Dept. of Telecommunication Engineering Fee Ctu Prague Czech Republic

ISBN: (纸本)9781450396851

Supervised learning of classifiers for text classification and sentiment analysis relies on the availability of labels that may be either difficult or expensive to obtain. A standard procedure is to add labels to the training dataset sequentially by querying an annotator until the model reaches a satisfactory performance. Active learning is a process that optimizes unlabeled data records selection for which the knowledge of the label would bring the highest discriminability of the dataset. Batch active learning is a generalization of a single instance active learning by selecting a batch of documents for labeling. This task is much more demanding because plenty of different factors come into consideration (i. e. batch size, batch evaluation, etc.). In this paper, we provide a large scale study by decomposing the existing algorithms into building blocks and systematically comparing meaningful combinations of these blocks with a subsequent evaluation on different text datasets. While each block is known (warm start weights initialization, Dropout MC, entropy sampling, etc.), many of their combinations like Bayesian strategies with agglomerative clustering are first proposed in our paper with excellent performance. Particularly, our extension of the warm start method to batch active learning is among the top performing strategies on all datasets. We studied the effect of this proposal comparing the outcomes of varying distinct factors of an active learning algorithm. Some of these factors include initialization of the algorithm, uncertainty representation, acquisition function, and batch selection strategy. Further, various combinations of these are tested on selected NLP problems with documents encoded using RoBERTa embeddings. Datasets cover context integrity (Gibberish Wackerow), fake news detection (Kaggle Fake News Detection), categorization of short texts by emotional context (Twitter Sentiment140), and sentiment classification (Amazon Reviews). Ultimately, we sh

关键词： Entropy

来源：评论

学校读者我要写书评

暂无评论

Proposals for the Promotion of Computing in K-12 Studies in Spain 53

Proposals for the Promotion of Computing in K-12 Studies in ...

引用

53rd IEEE ASEE Frontiers in Education International Conference, FIE 2023

作者： Lopez, David Llorens-Largo, Faraón Marqués-Andrés, Mercedes Velázquez-Iturbide, J. Ańgel Universitat Politècnica de Catalunya Computer Architecture Department Barcelona Spain Intelligence University of Alicante Dept. Computer Science and Artificial Alicante Spain University Jaume I Dept. Computer Science and Engineering Castelló Spain Universidad Rey Juan Carlos Departamento de Informática y Estadística Madrid Spain

ISBN: (纸本)9798350336429

The information society requires new skills and knowledge about how to cope in a computerized world. Although it is fully assumed in the political and social discourse, it is a neglected subject in formal education. Each nation should analyze how computing is learned in its K-12 education system. This research-to-practice paper presents the report carried out in Spain promoted by two academic associations, which analyzes the state of computing education in Spain in K-12 studies, summarizes some elements of the didactics of computing and analyzes the training received by teachers who will train K-12 students. The report also presents some reflections of the working group in charge, and nine proposals for improvement oriented to policy makers are pointed out. © 2023 IEEE.

关键词： Computing Pedagogy Culture of Teaching and Learning K12 STEM Education Teacher Development

来源：评论

学校读者我要写书评

暂无评论

2D LiDAR-Based Eucalyptus Tree Measurement and Monitoring System using Mobile Application 13

2D LiDAR-Based Eucalyptus Tree Measurement and Monitoring Sy...

引用

13th International Electrical engineering Congress, iEECON 2025

作者： Vanna-Iampikul, Pruek Boonnateephisit, Visutr Limsoontharakul, Somphop Moonrinta, Jednipat Dailey, Matthew Ekpanyapong, Mongkol Burapha University Department of Electrical Engineering Chonburi Thailand School of Engineering and Technology Asian Institute of Technology Pathumthani Thailand Prince of Songkla University Dept. Computational Science Songkhla Thailand Computer Science Field of Study School of Engineering and Technology Asian Institute of Technology Pathumthani Thailand Information & Communication Tech. School of Engineering and Technology Asian Institute of Technology Pathumthani Thailand School of Engineering and Technology Asian Institute of Technology Dept. of Industrial Systems Engineering Pathumthani Thailand

ISBN: (纸本)9798331543952

LiDAR technology has rapidly advanced, offering highly accurate solutions for tree geometry estimation, which is particularly valuable in agriculture and forestry. Its ability to capture precise 2D and 3D data has revolutionized tree detection and measurement. In Eucalyptus plantations, continuous monitoring is critical to maximizing growth, enhancing paper production, and mitigating environmental damage. However, existing solutions face challenges in real-time monitoring due to factors such as parasitic interference, low-light conditions, and uneven terrain. In this work, we present a novel, end-to-end framework that enhances Eucalyptus tree measurement and monitoring. Our framework significantly improves data collection efficiency by allowing measurements at walking speed, reducing measurement distance by 50% compared to traditional manual approaches. This system integrates 2D-LiDAR with additional environmental sensors to extract key tree parameters, including circumference, position, and tree count. The collected data is stored in real-time and uploaded to a centralized server for further analysis. Experimental evaluations across 160 trees, under five distinct farm conditions, demonstrate that our system achieves 89% accuracy in tree diameter estimation and 98.5% accuracy in tree counting, significantly outperforming traditional methods. These results establish our framework as a highly effective solution for improving the accuracy and efficiency of tree monitoring in challenging environments. © 2025 IEEE.

关键词： Deforestation

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：