检索结果-内蒙古大学图书馆

DeepEarthMY: A Remote Sensing Dataset for Tropical Land-Cover Segmentation

IEEE Access 2025年 13卷 69103-69115页

作者： Mohammed Najib, Shaaban Pee, Chih-Yang Ting, Choo-Yee Wong, Lai-Kuan Multimedia University Faculty of Computing and Informatics Selangor Cyberjaya63100 Malaysia

Land-cover mapping is essential for applications such as urban planning, natural resource management, and environmental monitoring. However, tropical equatorial regions pose unique challenges to land-cover classification due to dense vegetation, persistent cloud cover, and spectral similarity across land-cover types. Moreover, existing land-cover classification models often perform poorly in these regions as they are primarily trained on datasets from non-equatorial climates. To address this gap, we introduce DeepEarthMY, a high-resolution land-cover mapping dataset specifically designed for equatorial regions. The dataset contains 4,007 meticulously annotated images from 52 diverse regions across Malaysia, representing key land-cover types such as forests, buildings, roads, water bodies, agricultural lands, and barren land. We evaluated the dataset by performing land-cover segmentation on selected state-of-the-art semantic segmentation models, including DC-Swin, HRNet, and UNetFormer. The experimental results reveal that the DC-Swin model achieves the best performance, with a mIoU score of 80.23%. To further evaluate the significance of region-specific datasets for land-cover classification, we performed cross-dataset testing on two datasets: DeepEarthMY from the equatorial region and LoveDA from the temperate climate, using five-fold cross-validation. HRNet models trained on DeepEarthMY achieved a mIoU of 77.2% on DeepEarthMY but only 33.2% on the LoveDA test set. Conversely, models trained on LoveDA achieved 51.4% and 43% on the LoveDA and DeepEarthMY test sets, respectively. This significant performance gap highlights the need for region-specific datasets to advance land-cover research, particularly in equatorial climates where land-cover datasets are scarce. In conclusion, the DeepEarthMY dataset can be an invaluable resource for remote sensing in the equatorial region. The dataset is available at https://***/records/14242124. © 2013 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

An Integrated Web-Based Approach for Email Analysis Investigation Tool 24

An Integrated Web-Based Approach for Email Analysis Investig...

引用

9th International Conference on Cloud computing and Internet of Things, CCIOT 2024

作者： Bari, Fawzul Rahman, Siti Husna Abdul Faculty of Computing & Informatics Multimedia University Selangor Malaysia

ISBN: (纸本)9798400717161

Phishing websites have often been used for online and email fraud, spreading unwanted content. It has been important to find them before any user data was entered. In recent years, numerous attempts have been made to identify these phishing websites. To identify phishing web pages, most current methods trained classification models using manually constructed lexical and statistical features from the textual content of websites. The time-consuming method of extracting hand-crafted features, which required specialized domain knowledge to determine which features were useful for a given platform, was one of the limitations of these phishing detection approaches. To address these issues, this project aimed to develop a website which is an Integrated Web-based Approach for Email Analysis Investigation Tool (IWAEAIT) that used an external API like Virus total to analyses phishing websites and ip-api to get the location of the email. The project's goal was to ensure that anyone could use this website to identify if an email was phishing or not, even if they had no prior IT experience. This project also included features like file attachment analysis and a feature that let users take a screenshot of the website without really visiting it to stop users from visiting phishing websites, URL scan analysis and email artefacts extractor to get the location of the email. Copyright © 2024 held by the owner/author(s).

关键词： Phishing

来源：评论

学校读者我要写书评

暂无评论

Analyzing Cybersecurity Risk with a Phishing Simulation Website 24

Analyzing Cybersecurity Risk with a Phishing Simulation Webs...

引用

9th International Conference on Cloud computing and Internet of Things, CCIOT 2024

作者： Ali, Alya Nasuha Binti Mohamed Rahman, Siti Husna Abdul Faculty of Computing & Informatics Multimedia University Selangor Malaysia

ISBN: (纸本)9798400717161

Phishing is a cyberattack that exploits human psychology and technical vulnerabilities. Previous research has revealed inconsistency of relation between age and phishing susceptibility. Hence, this project aims to identify which age group is more susceptible to phishing. Therefore, to achieve these objectives, this project developed and deployed a phishing simulation website to analyse which age group is more vulnerable. Respondents are considered successfully phished if they submitted information to the phishing website. The project adopted previous researcher's phishing process with addition of two phases as a guideline. The phases were Planning, Preparation, Attack, Valuable Acquisition, Knowledge Transfer and Result Analysis. A blog was developed to redirect phishing victims to an e-Learning page to raise their cybersecurity awareness. Persuasion techniques used in this simulation were Reward and Urgency. The phishing website gained 146 clicks and 89 form submissions. Adults aged 18 to 59 years old (group A) has a higher phishing success rate of compared to adults aged 60 years and above (group B). These findings suggest younger adults are more susceptible to phishing. Future research should involve larger sample size and seek broader input from participants. Copyright © 2024 held by the owner/author(s). Publication rights licensed to ACM.

关键词： Phishing

来源：评论

学校读者我要写书评

暂无评论

Mix-YOLONet: Deep Image Dehazing for Improving Object Detection 31st

Mix-YOLONet: Deep Image Dehazing for Improving Object Detec...

引用

31st International Conference on multimedia Modeling, MMM 2025

作者： Lim, Xin Wong, Lai-Kuan Loh, Yuen Peng Gu, Ke Lin, Weisi Faculty of Computing and Informatics Multimedia University Cyberjaya Malaysia Beijing University of Technology Beijing China Nanyang Technological University Singapore Singapore

ISBN: (纸本)9789819620609

Atmospheric haze significantly impairs the performance of computer vision tasks such as image dehazing and object detection. Existing methods often address these tasks independently, failing to provide an integrated solution that can effectively handle hazy conditions while maintaining accurate object detection. This paper presents Mix-YOLONet, a novel joint network architecture designed to tackle both image dehazing and object detection simultaneously. Mix-YOLONet leverages the powerful image restoration capabilities of U-Net-like architecture and integrates it with the detection precision of a YOLO head. Additionally, we integrated Mix Structure Blocks (MSB) to our joint network and experimented various configurations at strategic locations to enhance feature extraction and context aggregation. Through extensive experiments, we demonstrate that Mix-YOLONet achieves superior performance in both dehazing and object detection tasks under challenging visibility conditions, outperforming state-of-the-art methods on three benchmark datasets both quantitatively and qualitatively. The proposed joint network not only improves the clarity of hazy images but also ensures accurate object localization, paving the way for more robust object detection in adverse environmental conditions. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Restoration

来源：评论

学校读者我要写书评

暂无评论

Leveraging LLMs for optimised feature selection and embedding in structured data: A case study on graduate employment classification

引用

Computers and Education: Artificial Intelligence 2025年 8卷

作者： Haque, Radiah Goh, Hui-Ngo Ting, Choo-Yee Quek, Albert Hasan, M.D. Rakibul Faculty of Computing and Informatics Multimedia University Cyberjaya 63100 Malaysia

The application of Machine Learning (ML) for predicting graduate student employability is a growing area of research, driven by the need to align educational outcomes with job market requirements. In this context, this paper investigates the application of Large Language Models (LLMs) for tabular data transformation and embedding, specifically using Bidirectional Encoder Representations from Transformers (BERT), to enhance the performance of ML models in binary classification tasks for student employability prediction. The primary objective is to determine whether converting structured data into text format improves model accuracy. The study involves several ML models including Artificial Neural Networks (ANN), CatBoost, and BERT classifier. The focus is on predicting the employment status of graduate students based on demographic, academic, and graduate tracer study data, collected from over 4000 university graduates. Feature selection methods, including Boruta and Extra Tree Classifier (ETC) are employed to identify the optimal feature set, guided by a sliding window algorithm for automatic feature selection. The models are trained in four stages: 1) original dataset without feature selection or word embedding, 2) dataset with selected optimal features, 3) transformed data with word embedding, and 4) transformed data with feature selection applied both before and after word embedding. The baseline model (without feature selection and embedding) achieved the highest accuracy with the ANN model (79%). Subsequently, applying ETC for feature selection improved accuracy, with CatBoost achieving 83%. Further transformation with BERT-based embeddings raised the highest accuracy to 85% using the BERT classifier. Finally, the optimal accuracy of 88% was obtained by applying feature selection before and after embedding, with the BERT-Boruta model. The findings from this study demonstrate that using the dual-stage feature selection approach in combination with BERT embedding

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Machine Learning-Driven QoT Prediction for Enhanced Optical Networks in DWDM system

引用

IEEE Access 2025年 13卷 80445-80460页

作者： Sakthivel, Sudha Belgaum, Mohammad Riyaz Sajak, Aznida Abu Bakar Alam, Muhammad Mansoor Su'ud, Mazliham Mohd Universiti Kuala Lumpur Advanced Telecommunication Technology Research Cluster Malaysia University of Technology and Applied Sciences Muscat133 Oman Multimedia University Faculty of Computing and Informatics Cyberjaya63100 Malaysia Arab Open University Faculty of Computer Studies 18211 Bahrain Riphah International University Faculty of Computing Islamabad46000 Pakistan

This research demonstrated the machine learning (ML) classifiers with regression learning to improve an optical system’s quality of transmission (QoT). In Optical Communication, the data can be communicated from source to destination through the established lightpaths. However, as the signal traverses the optical links and devices, its QoT may deteriorate due to various impairments. The QoT is an essential component that determines the connectivity of an optical network. Therefore, ensuring a QoT guarantee is necessary to establish a successful lightpath. Predicting the QoT before establishing lightpaths can guide the routing and allocation of resources required for the lightpaths. In this research, using ML models an appropriate QoT analytical prediction model is developed computationally. Simulations were conducted at a 10 Gbps data rate per channel for 64-channel DWDM systems. The proposed model significantly improves in detecting fiber nonlinearity, and performance was studied using Q-factor, BER, and noise power. The results indicate that the SVM-based classifier with regression learning performs better than any other classifiers discussed in this research. This study assesses the efficiency of the proposed ML models in predicting the QoT for established lightpaths. Results indicate that all the ML classifiers with Regression models can accurately predict the transmission quality for over 90% of lightpaths. However, the proposed SVM-based classifier with a regression model demonstrates superior generalization, with a nearly perfect QoT prediction rate of around 99% for the established lightpaths. In the network planning stage, residual margins are added to compensate for inaccuracies, which ensures accurate signal reception. The proposed ML model achieved a lightpath residual margin with a 0.7 dB error. © 2013 IEEE.

关键词： Fiber to the x

来源：评论

学校读者我要写书评

暂无评论

M3-Net: A Multi-Scale Nuclei Segmentation Model for Breast Cancer Histopathology Using Contextual Patches and Attention Mechanism

M3-Net: A Multi-Scale Nuclei Segmentation Model for Breast C...

引用

IEEE International Symposium on Biomedical Imaging

作者： Arbab Sufyan Mohammad Faizal Ahmad Fauzi Wong Lai Kuan Faculty of Engineering Multimedia University Cyberjaya Malaysia Faculty of Computing & Informatics Multimedia University Cyberjaya Malaysia

ISBN: (数字)9798331520526

ISBN: (纸本)9798331520533

Histopathological analysis, particularly of nuclear morphology, is critical for identifying malignancies. Accurate nuclei segmentation plays a pivotal role in this process, as it enables detailed assessment of nuclear size, shape, and distribution patterns. Traditional segmentation methods, however, often fail to capture fine details, lack broader context, and struggle with overlapping nuclei of varying sizes and shapes, especially when relying on single-scale approaches. To address these challenges, we propose a multi-scale context-based encoder-decoder model named M3-Net (Multi-Scale, Multi-Level, Multi-Stream Network) that integrates both global and local tissue features. Evaluations demonstrate that M3-Net effectively segments overlapping nuclei and diverse structures, providing a robust solution for automated nuclei segmentation in breast cancer pathology.

关键词： Image segmentation Accuracy Image analysis Histopathology Shape Image color analysis Morphology Breast cancer Context modeling Biomedical imaging

来源：评论

学校读者我要写书评

暂无评论

ERBM: A Machine Learning-Driven Rule-Based Model for Intrusion Detection in IoT Environments

引用

Computers, Materials and Continua 2025年第3期83.0卷 5155-5179页

作者： Mehmmod, Arshad Batool, Komal Sajid, Ahthsham Alam, Muhammad Mansoor Khan, Inam Ullah Department of Information Security and Data Science Riphah Institute of Systems Engineering Riphah International University Islamabad 46000 Pakistan Faculty of Computing Riphah International University Islamabad 46000 Pakistan Faculty of Computing and Informatics Multimedia University Cyberjaya 63100 Malaysia

Traditional rule-based Intrusion Detection Systems (IDS) are commonly employed owing to their simple design and ability to detect known threats. Nevertheless, as dynamic network traffic and a new degree of threats exist in IoT environments, these systems do not perform well and have elevated false positive rates—consequently decreasing detection accuracy. In this study, we try to overcome these restrictions by employing fuzzy logic and machine learning to develop an Enhanced Rule-Based Model (ERBM) to classify the packets better and identify intrusions. The ERBM developed for this approach improves data preprocessing and feature selections by utilizing fuzzy logic, where three membership functions are created to classify all the network traffic features as low, medium, or high to remain situationally aware of the environment. Such fuzzy logic sets produce adaptive detection rules by reducing data uncertainty. Also, for further classification, machine learning classifiers such as Decision Tree (DT), Random Forest (RF), and Neural Networks (NN) learn complex ways of attacks and make the detection process more precise. A thorough performance evaluation using different metrics, including accuracy, precision, recall, F1 Score, detection rate, and false-positive rate, verifies the supremacy of ERBM over classical IDS. Under extensive experiments, the ERBM enables a remarkable detection rate of 99% with considerably fewer false positives than the conventional models. Integrating the ability for uncertain reasoning with fuzzy logic and an adaptable component via machine learning solutions, the ERBM system provides a unique, scalable, data-driven approach to IoT intrusion detection. This research presents a major enhancement initiative in the context of rule-based IDS, introducing improvements in accuracy to evolving IoT threats. Copyright © 2025 The Authors.

关键词： fuzzy prediction intrusions IoT Rule based

来源：评论

学校读者我要写书评

暂无评论

Face image authentication scheme based on MTCNN and SLT

引用

multimedia Tools and Applications 2025年 1-43页

作者： Thabit, Rasha Al-Askari, Mohanad A. Mohammed, Dunya Zeki Anaam, Elham Abdulwahab Mahmood, Zainab H. Jabbar, Dina Jamal Salih, Zahraa Aqeel Information System Department College of Computer Sciences and Information Technology University of Al-Anbar Al-Anbar Iraq Computer Engineering Department College of Engineering Al-Iraqia University Baghdad Iraq Electronic and Communications Engineering Department College of Engineering Gilgamesh University Baghdad Iraq Faculty of Computing and Informatics Multimedia University Selangor Cyberjaya Malaysia Department of Cybersecurity Engineering College of Information Engineering Al-Nahrain University Baghdad Iraq Electrical Engineering Technical Collage Middle Technical University Baghdad Iraq Department of Business Administration Alamaal University College Baghdad Iraq

DeepFakes and face image manipulation methods have been widely distributed in the last few years and several techniques have been presented to check the authenticity of the face image and detect manipulation if exists. Most of the available manipulation detection techniques have been successfully applied to reveal one type of manipulation under specific conditions, however, many limitations and challenges can be encountered in this field. To overcome some limitations and challenges, this paper presents a new face image authentication (FIA) scheme based on Multi-Task Cascaded Conventional Neural Networks (MTCNN) and watermarking in Slantlet transform (SLT) domain. The proposed FIA scheme has three main algorithms that are face detection and selection, embedding, and extraction algorithms. Different block sizes have been used to divide the image into non-overlapping blocks followed by classifying them into two groups that are blocks from face area (FA) and blocks from the remaining area (RA) of the image. In the embedding algorithms, the authentication information is generated from FA blocks and embedded in the RA blocks. In the extraction algorithms, the embedded information is extracted from RA blocks and compared with the calculated data from FA blocks to reveal manipulations and localize the manipulated blocks if exist. Extensive experiments have been conducted to evaluate the performance of the proposed FIA scheme for different face images. The experimental work included tests for payload, capacity, visual quality, time complexity, and localization of manipulations. The results proved the efficiency of the proposed scheme in detecting and localizing different face image manipulations such as attributes attacks, retouching attacks, expression swap, face swap, and morphing attacks. The proposed scheme overcomes many limitations and it is 100% accurate in localizing the tampered blocks which makes it a better candidate for practical applications. © The Author(s), un

关键词： Authentication

来源：评论

学校读者我要写书评

暂无评论

Financial Data-Driven Classification of Manufacturing Companies: A Case Study of the Malaysian Industry 3rd

Financial Data-Driven Classification of Manufacturing Compan...

引用

3rd International Conference on Advancements in Smart computing and Information Security, ASCIS 2024

作者： Kannan, Rathimala Ramakrishnan, Kannan PSG Institute of Management PSG College of Technology Tamil Nadu Coimbatore641004 India Faculty of Computing and Informatics Multimedia University Selangor Cyberjaya63100 Malaysia

ISBN: (纸本)9783031863011

Manufacturing sector companies are typically categorized based on size or product type to aid in policy formulation and long-term planning. The existing classification systems, however, overlook economic performance, which makes it difficult to plan policies precisely and allocate resources. This study aims to close this gap by identifying economic factors that set manufacturing companies apart using data mining and machine learning. The main objective is to uncover similarity patterns among manufacturing companies and to describe each segment’s characteristics addressed in this study by utilizing historical data from the Manufacturing Economic Census Survey. The Cross-Industry Standard Process for Data Mining (CRISP-DM) framework guides the extraction of meaningful patterns from the economic census survey data. The study attempts to efficiently group manufacturing organizations according to their economic performance by utilizing unsupervised learning algorithms such as fuzzy C-means clustering and K-means clustering. The findings deviate from the existing categorizing method and yield six groups that are defined by financial metrics such as net book value, staff qualities, and spending distribution. Each segment is described in detail, including key characteristics such as business structure, primary activities, workforce qualifications, and financial metrics. These insights are instrumental in making informed decisions regarding future policy changes and financial aid initiatives, ultimately fostering a more resilient and competitive manufacturing sector. This research underscores the significance of data-driven approaches in elucidating industry dynamics and guiding strategic interventions to propel economic growth and enhance the well-being of Malaysians. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Competition

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：