检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Balat, Mazen Awaad, Rewaa Zaky, Ahmed B. Aly, Salah A. CS & IT Department Egypt-Japanese University of Science & Technology Alexandria Egypt Faculty of Computing and Data Science Badya University Giza Egypt Computer Science Section Faculty of Science Fayoum University Fayoum Egypt

This study introduces an integrated approach to recognizing Arabic Sign Language (ArSL) using state-of-the-art deep learning models such as MobileNetV3, ResNet50, and EfficientNet-B2. These models are further enhanced by explainable AI (XAI) techniques to boost interpretability. The ArSL2018 and RGB Arabic Alphabets Sign Language (AASL) datasets are employed, with EfficientNet-B2 achieving peak accuracies of 99.48% and 98.99%, respectively. Key innovations include sophisticated data augmentation methods to mitigate class imbalance, implementation of stratified 5-fold cross-validation for better generalization, and the use of Grad-CAM for clear model decision transparency. The proposed system not only sets new benchmarks in recognition accuracy but also emphasizes interpretability, making it suitable for applications in healthcare, education, and inclusive communication technologies. © 2025, CC0.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Modified simulated annealing for university teacher course assignment considering socio-cultural constraints

引用

International Journal of Information Technology (Singapore) 2025年第3期17卷 1533-1549页

作者： Brahimi, Samiha Sabba, Sara Elhussein, Mariam Alqahtani, Mohammed Department of Computer Information Systems College of Computer Science and Information Technology Imam Abdulrahman Bin Faisal University Dammam Saudi Arabia Faculty of New Technologies of Information and Communication Laboratory of Data Science and Artificial Intelligence (LISIA) Abdelhamid Mahri Constantine-2 University Constantine 25016 Algeria Department of Mathematics and Computing Mount Royal University Calgary AB Canada

This paper aims to address the complex challenge of course assignment for faculty members within a Saudi university, taking into account the socio-cultural constraints imposed by gender-based segregation between students and faculty. To tackle this challenge, a combinatorial optimization model is proposed, which includes a binary representation of the problem, constraints reflecting gender segregation, and an objective function designed to maximize faculty satisfaction. The model employs a Simulated Annealing (SA) algorithm, supplemented with two neighborhood search heuristics (H1 and H2), to explore diverse combinations. Additionally, a greedy randomized adaptive search procedure Greedy Randomized Adaptive Search Procedure (GRASP) algorithm is implemented for comparative analysis, utilizing heuristic 1 for local search. Through the application of simulated annealing alongside both heuristics (SA_H1_H2), the optimal solution is achieved, resulting in a notable equilibrium of courses among faculty members. Objective 1 demonstrates the standardized deviation in workload distribution among faculty members, yielding a deviation value of 0.08. This study contributes to the field by proposing a novel approach to course assignment that explicitly addresses the gender-based segregation prevalent in Saudi universities. By integrating socio-cultural constraints into the optimization model and employing advanced algorithms, the study offers a pioneering solution to a previously unexplored problem domain. © Bharati Vidyapeeth's Institute of Computer Applications and Management 2024.

关键词： GRASP Mathematical model Operations Research Optimization Simulated annealing Teacher_Course assignment

来源：评论

学校读者我要写书评

暂无评论

AI Driven Blockchain Reward Insights Using ML

AI Driven Blockchain Reward Insights Using ML

引用

Emerging Trends in Electrical, Control, and Telecommunication Engineering (ETECTE), International Conference on

作者： Abeera Malik Faisal Bukhari Hafiz Muhammad Talha Department of Information Technology Faculty of Computing and Information Technology University of the Punjab PUCIT Lahore Pakistan Department of Data Science Faculty of Computing and Information Technology University of the Punjab Lahore Pakistan

ISBN: (数字)9798331504861

ISBN: (纸本)9798331504878

Several industries have embraced blockchain technology due to the change it has brought, especially in the way records are kept. However, one major problem inherent in the concept of blockchain from the participants perspective is the presence of unpredictable rewards in proof-of-work and proof- of-stake models. That is why this research aims at developing a new strategy to solve this problem by incorporating artificial intelligence (AI) as well as machine learning (ML) algorithms into blockchain compensation prediction. This project seeks to analyze how artificial intelligence (AI) and machine learning (ML) methods may be incorporated into blockchain technology to solve the problem of the unpredictability of rewards. Using the EtherScan API, we have obtained 1000 timestamps and block reward values, and for prediction of rewards for certain timestamps, we use the KNN algorithm, linear regression algorithm, and random forest regressor algorithm. Thus, a What We KNN attained an accuracy of performance specifies, and a random forest regressor attained an impression accuracy range of 74% particular timestamps. These outcomes show the benefits of advanced ML to the blockchain ecosystem as a tool that helps stakeholders make informed decisions concerning activities like mining, staking, or investing in blockchains with potentially high rewards.

关键词： Technological innovation Machine learning algorithms Accuracy Biological system modeling Nearest neighbor methods Prediction algorithms Proof of Work Blockchains Stakeholders Random forests

来源：评论

学校读者我要写书评

暂无评论

Optimized Hybrid CNN-GRU Approach for Accurate Prediction of Cardiovascular Diseases: Integrating Medical Imaging and Clinical data 3

Optimized Hybrid CNN-GRU Approach for Accurate Prediction of...

引用

3rd International Conference on Communication, Control, and Intelligent Systems, CCIS 2024

作者： Prasanna, V. Suresh Kumar, N. Divyaprabha, P. Sundaram, R. Srm Institute of Science and Technology School of Computing Faculty of Engineering and Technology Department of Data Science and Business Systems Chennai India Sri Ramakrishna Engineering College Department of Information Technology Coimbatore India Sri Ramakrishna Engineering College Department of Computer Science and Engineering Coimbatore India Ganesh College of Engineering Department of Computer Science and Engineering Salem India

ISBN: (纸本)9798331528201

As the primary cause of death globally, cardiovascular diseases (CVDs) demand precise and timely prediction to enhance patient outcomes. Other examples of conventional approaches for CVD prediction include statistical models as well as machine learning models such as a single model that often struggle to address complex and comprehensive data from clinical records and Medical Imaging. To address these issues, the concept proposed here is to integrate the medical imaging data with clinical data as new apart from utilizing the hybrid CNN-GRU model to predict cardiovascular diseases with higher accuracy and interpretability. While the GRU module deals with the time sequences of clinical data including past history and other test results, the CNN module is to extract the spatial features from the medical images. Thus, by integrating these modules, it is more accurate to predict the data of the patient as a synergy of spatial and temporal arrangements. Cross-validation techniques and performance indices are used for model training and testing on a diverse data set containing imaging and other clinical-related data. Experimental results depict that the suggested hybrid CNN-GRU model performs much better than the traditional approaches and has higher levels of predicted accuracy and robustness. It is also feasible to generalize this integrated strategy to other medical disorder where multi-modal data analysis is crucial, besides enhancing the capacity to estimate CVDs. The findings underline the importance of integrating doctors' information with radiological images and applying new patterns of neural networks to augmenting clinical judgment. The suggested work is implemented using python. © 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

A Novel 3D Medical Image Segmentation Model Using Improved SAM

A Novel 3D Medical Image Segmentation Model Using Improved S...

引用

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

作者： Kuang, Yuansen Ma, Xitong Zhao, Jing Wang, Guangchen Zeng, Yijie Liu, Song Key Laboratory of Computing Power Network and Information Security Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China

ISBN: (纸本)9781665410205

3D medical image segmentation is an essential task in the medical image field, which aims to segment organs or tumours into different labels. A number of issues exist with the current 3D medical image segmentation task: existing models cannot simultaneously obtain the space correlation and depth correlation of 3D slices;previous models suffer from local detail loss of positional embedding in 3D images;previous approaches often have blurring of boundaries in segmenting 3D images. To solve these shortcomings, we propose a 3D medical image segmentation model named TPM-SAM. In our model, we design a twinchannel image encoder to simultaneously capture the space correlation and depth correlation of 3D slices through a multi-head attention mechanism and improved adapters. Furthermore, we design a prompt encoding generator, which divides the volumetric image into small blocks and better captures the local detail information. In addition, we introduce a multi-layer aggregation decoder by employing U-Net with multi-level skip connection to solve the blurring of boundaries in processing 3D images. Finally, we experimented and evaluated our model on KiTS21 and LiTS17 datasets to compare with other baseline models. © 2024 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

The Evolution of Zero Trust Architecture (ZTA) from Concept to Implementation

arXiv

引用

arXiv 2025年

作者： Nasiruzzaman, Md. Ali, Maaruf Salam, Iftekhar Miraz, Mahdi H. School of Computing & Data Science Xiamen University Malaysia Malaysia Faculty of Computer Science IT Universiteti Metropolitan Tirana Albania The Doctoral College University of Wales Trinity Saint David Wales United Kingdom School of Computing Faculty of Arts Science & Technology Wrexham University United Kingdom Faculty of Computing Engineering & Science University of South Wales United Kingdom

Zero Trust Architecture (ZTA) is one of the paradigm changes in cybersecurity, from the traditional perimeter-based model to perimeterless. This article studies the core concepts of ZTA, its beginning, a few use cases and future trends. Emphasising the always-verify and least privilege access, some key tenets of ZTA have grown to be integration technologies like Identity Management, Multi-Factor Authentication (MFA) and real-time analytics. ZTA is expected to strengthen cloud environments, education, work environments (including from home) while controlling other risks like lateral movement and insider threats. Despite ZTA’s benefits, it comes with challenges in the form of complexity, performance overhead and vulnerabilities in the control plane. These require phased implementation and continuous refinement to keep up with evolving organisational needs and threat landscapes. Emerging technologies, such as Artificial Intelligence (AI) and Machine Learning (ML) will further automate policy enforcement and threat detection in keeping up with dynamic cyber threats. Copyright © 2025, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Enhancing data Quality in Image Pre-processing: A Case Study on Plant Disease Classification

Enhancing Data Quality in Image Pre-processing: A Case Study...

引用

2023 International Conference for Technological Engineering and its Applications in Sustainable Development, ICTEASD 2023

作者： Ong, Ge Ye Abdullah, Nurzulaikha Ridzuan, Fakhitah Mustafa, Wan Azani Alzubaidi, Laith H. INTI International College Penang School of Engineering and Technology Penang Malaysia Universiti Malaysia Kelantan Faculty of Data Science and Computing Kota Bharu Malaysia Universiti Malaysia Perlis Faculty of Electrical Engineering & Technology Perlis Arau02600 Malaysia The Islamic University Najaf Faculty of Engineering Iraq

ISBN: (纸本)9798350336474

data quality plays a vital role in image pre-processing for machine learning applications. The effectiveness of building accurate and reliable models lies in the high-quality of data. Prioritizing data quality in the initial stages of the image pre-processing pipeline lays a strong foundation for subsequent machine learning stages. Therefore, this research aims to identify the appropriate steps for image pre-processing and compare the performance of machine learning model based on different pre-processing approaches. In this study, we utilized a plant diseases dataset sourced from Kaggle, comprising approximately 87,000 RGB images of both healthy and diseased crop leaves, categorized into 38 distinct classes. To significantly enhance the image quality, we implemented a range of transformation techniques, including resizing, normalization, data augmentation, and cropping. The analysis clearly indicates that implementing comprehensive pre-processing techniques enhances data quality and improves the machine learning model's classification performance. © 2023 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Generative AI for synthetic data across multiple medical modalities: A systematic review of recent developments and challenges

引用

Computers in Biology and Medicine 2025年 189卷 109834-109834页

作者： Ibrahim, Mahmoud Khalil, Yasmina Al Amirrajab, Sina Sun, Chang Breeuwer, Marcel Pluim, Josien Elen, Bart Ertaylan, Gökhan Dumontier, Michel Institute of Data Science Faculty of Science and Engineering Maastricht University Maastricht Netherlands Department of Advanced Computing Sciences Faculty of Science and Engineering Maastricht University Maastricht Netherlands VITO Belgium Department of Biomedical Engineering Eindhoven University of Technology Eindhoven Netherlands

This paper presents a comprehensive systematic review of generative models (GANs, VAEs, DMs, and LLMs) used to synthesize various medical data types, including imaging (dermoscopic, mammographic, ultrasound, CT, MRI, and X-ray), text, time-series, and tabular data (EHR). Unlike previous narrowly focused reviews, our study encompasses a broad array of medical data modalities and explores various generative models. Our aim is to offer insights into their current and future applications in medical research, particularly in the context of synthesis applications, generation techniques, and evaluation methods, as well as providing a GitHub repository as a dynamic resource for ongoing collaboration and innovation. Our search strategy queries databases such as Scopus, PubMed, and ArXiv, focusing on recent works from January 2021 to November 2023, excluding reviews and perspectives. This period emphasizes recent advancements beyond GANs, which have been extensively covered in previous reviews. The survey also emphasizes the aspect of conditional generation, which is not focused on in similar work. Key contributions include a broad, multi-modality scope that identifies cross-modality insights and opportunities unavailable in single-modality surveys. While core generative techniques are transferable, we find that synthesis methods often lack sufficient integration of patient-specific context, clinical knowledge, and modality-specific requirements tailored to the unique characteristics of medical data. Conditional models leveraging textual conditioning and multimodal synthesis remain underexplored but offer promising directions for innovation. Our findings are structured around three themes: (1) Synthesis applications, highlighting clinically valid synthesis applications and significant gaps in using synthetic data beyond augmentation, such as for validation and evaluation;(2) Generation techniques, identifying gaps in personalization and cross-modality innovation;and (3) Evalu

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

LESaET: Low-Dimensional Embedding Method for Link Prediction Combining Self-Attention and Enhanced-TuckER

LESaET: Low-Dimensional Embedding Method for Link Prediction...

引用

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

作者： Ding, Lichao Zhao, Jing Lu, Kai Hao, Zenghao Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (纸本)9781665410205

Knowledge graphs (KGs) provide a structured representation of the real world through entity-relation triples. However, current KGs are often incomplete, typically containing only a small fraction of all possible facts. This involves inferring missing content from existing information, a task known as link prediction. Existing methods in the field of link prediction struggle with controlling the dimensionality of embedding vectors or suffer from overly complex models. In order to tackle these challenges, we introduce a method in this paper, named Low-Dimensional Embedding Method for Link Prediction Combining Self-Attention and Enhanced-TuckER (LESaET). LESaET leverages both self-attention mechanisms and tensor factorization to learn expressive contextual-enhanced representations of KGs. Specifically, LESaET employs the multi-head self-attention mechanism of Transformer as an encoder to capture the mutual information between entities and relationships, and utilizes Enhanced-TuckER as a decoder, ultimately achieving expressive low-dimensional embeddings for link prediction tasks. LESaET demonstrates competitive performance compared to advanced methods on standard datasets. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Deep Fourier-embedded Network for RGB and Thermal Salient Object Detection

arXiv

引用

arXiv 2024年

作者： Lyu, Pengfei Yeung, Pak-Hei Yu, Xiaosheng Wu, Chengdong Rajapakse, Jagath C. Faculty of Robot Science and Engineering Northeastern University Shenyang110169 China College of Computing and Data Science Nanyang Technological University 639798 Singapore

The rapid development of deep learning has significantly improved salient object detection (SOD) combining both RGB and thermal (RGB-T) images. However, existing deep learning-based RGB-T SOD models suffer from two major limitations. First, Transformer-based models with quadratic complexity are computationally expensive and memory-intensive, limiting their application in high-resolution bi-modal feature fusion. Second, even when these models converge to an optimal solution, there remains a frequency gap between the prediction and ground-truth. To overcome these limitations, we propose a purely Fourier transform-based model, namely Deep Fourier-Embedded Network (DFENet), for accurate RGB-T SOD. To address the computational complexity when dealing with high-resolution images, we leverage the efficiency of fast Fourier transform with linear complexity to design three key components: (1) the Modal-coordinated Perception Attention, which fuses RGB and thermal modalities with enhanced multi-dimensional representation;(2) the Frequency-decomposed Edge-aware Block, which clarifies object edges by deeply decomposing and enhancing frequency components of low-level features;and (3) the Fourier Residual Channel Attention Block, which prioritizes high-frequency information while aligning channel-wise global relationships. To mitigate the frequency gap, we propose Co-focus Frequency Loss, which dynamically weights hard frequencies during edge frequency reconstruction by cross-referencing bi-modal edge information in the Fourier domain. Extensive experiments on four RGB-T SOD benchmark datasets demonstrate that DFENet outperforms fifteen existing state-of-the-art RGB-T SOD models. Comprehensive ablation studies further validate the value and effectiveness of our newly proposed components. The code is available at https://***/JoshuaLPF/DFENet. Copyright © 2024, The Authors. All rights reserved.

关键词： Object detection

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

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

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：