检索结果-内蒙古大学图书馆

2nd International Conference on Machine Learning and Autonomous Systems, ICMLAS 2025

作者： Ebadinezhad, Sahar Nia, Nooshin Nooraei Shirzad, Nasratullah Osemeha, Nwabueze Kenneth Near East University Department of Computer Information Systems Nicosia99138 Cyprus Artificial Intelligence Near East University Nicosia Cyprus Near East University Software Engineering Department Nicosia Cyprus

ISBN: (纸本)9798331505745

Due to The lack of comparison studies and practical applications of RNN, LSTM, and hybrid RNN-LSTM models for intrusion detection systems, especially when managing class imbalances in complex network datasets, represents a gap in the literature. This study investigates the application of deep learning techniques to improve the detection capabilities of IDS. For this study, we applied and assessed Recurrent Neural Network, Long Short-Term Memory, and Hybrid RNN-LSTM models on a UNSW-NB15 dataset, and to redress class imbalance we trained and validated our models using synthetic minority over-sampling. The RNN-LSTM model was shown to perform the best with a 9 4. 0 0 % accuracy in comparison to other models. These experimental results signify the possibility of enhancing IDS performance and a dependable method of detecting various network intrusions with the use of hybrid models comprised of RNN & LSTM. © 2025 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

ASSESSING AND PREDICTING AIR POLLUTION IN ASIA: A REGIONAL AND TEMPORAL STUDY (2018-2023)

arXiv

引用

arXiv 2025年

作者： Rahman, Anika Khatun, Taskia Department of Computer Science and Engineering Stamford University Bangladesh Dhaka Bangladesh Department of Software Engineering Daffodil International University Dhaka Bangladesh

This study analyzes and predicts air pollution in Asia, focusing on PM 2.5 levels from 2018 to 2023 across five regions: Central, East, South, Southeast, and West Asia. South Asia emerged as the most polluted region, with Bangladesh, India, and Pakistan consistently having the highest PM 2.5 levels and death rates, especially in Nepal, Pakistan, and India. East Asia showed the lowest pollution levels. K-means clustering categorized countries into high, moderate, and low pollution groups. The ARIMA model effectively predicted 2023 PM 2.5 levels (MAE: 3.99, MSE: 33.80, RMSE: 5.81, R : 0.86). The findings emphasize the need for targeted interventions to address severe pollution and health risks in South Asia. © 2025, CC BY.

关键词： Haze pollution

来源：评论

学校读者我要写书评

暂无评论

Exploiting Wavelet Scattering Transform & Squeeze-Excitation Blocks with Cross-Modal Attention for Multi-modal Emotion Recognition

Exploiting Wavelet Scattering Transform & Squeeze-Excitation...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Junchen Liu Jesin James Karan Nathwani Department of Electrical Computer and Software Engineering The University of Auckland New Zealand Department of Electrical Engineering India Institute of Technology Jammu India

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Multi-modal emotion recognition (MER) is crucial for improving human-computer interaction. Convolutional neural networks (CNNs) are the mainstream for MER tasks, but they require large databases, extensive memory, and significant energy, limiting their practical use. This paper proposes a novel MER system that leverages wavelet scattering transform (WST) to address these challenges, achieving improved performance with lower computational consumption. Moreover, the system benefits from the noise robustness provided by WST. By integrating WST as a non-trainable initial layer in a CNN model and employing an encoder module, our system effectively captures time-frequency, local and high-level features from both speech and video. We enhance feature integration and representation with cross-modal attention (CMA) and a squeeze-and-excitation (SE) block. The results demonstrate that our system performs consistently across varying noise levels and duration thresholds. Ablation studies reveal that the combination of MFCC, Mel spectrogram, and raw waveform features yields the highest accuracy, with Mel spectrogram being the most influential. Experimental results on the IEMOCAP and RAVDESS databases achieve emotion recognition accuracy of 83.2% and 97.8%, respectively, showcasing improved performance and robustness compared to state-of-the-art models, while using fewer trainable parameters.

关键词： Wavelet transforms Emotion recognition Time-frequency analysis Accuracy Databases Computational modeling Scattering Feature extraction Convolutional neural networks Spectrogram

来源：评论

学校读者我要写书评

暂无评论

WQuatNet: Wide range quaternion-based head pose estimation

引用

Journal of King Saud University - computer and Information Sciences 2025年第3期37卷 1-14页

作者： Algabri, Redhwan Shin, Hyunsoo Abdu, Ahmed Bae, Ji-Hun Lee, Sungon Dept. of Computer Science & Engineering College of AI Convergence Sejong University Seoul 05006 South Korea Research Institute of Engineering and Technology Hanyang University Ansan 15588 South Korea School of Software Northwestern Polytechnical University Xi’an 710072 China Department of Robotics Hanyang University Ansan 15588 South Korea

Head pose estimation (HPE) is a critical task for numerous applications ranging from human-computer interaction, healthcare, and robotics, to surveillance. Most existing methods employ Euler angles as a representation, which often face challenges such as a gimbal lock, especially in full-range rotation scenarios or rotation matrices that require nine parameters. This study introduces WQuatNet, a novel deep learning-based model that leverages the quaternion representation, which uses only four parameters, to avoid this challenge. WQuatNet was designed based on a landmark-free HPE method to predict head poses across the full-range angles of 360∘ from images. Landmark-free methods bypass the need for explicit detection of facial landmarks;instead, they leverage the entire image to estimate the head orientation. The model incorporates a RepVGG-D2se backbone for robust feature extraction and introduces two loss functions tailored for quaternion predictions. Our experimental results on multiple HPE datasets covering both narrow- and full-range angles demonstrate that WQuatNet outperforms the state-of-the-art (SOTA) approaches in terms of accuracy. The performance of the proposed HPE was evaluated using the CMU, AGORA, BIWI, AFLW2000, and 300W-LP datasets. We also perform ablation studies and error analyses to validate the significance of each component of the model. © The Author(s) 2025.

关键词： Deep neural network Full range of rotation Head pose estimation Quaternion

来源：评论

学校读者我要写书评

暂无评论

Evaluating Different Hybrid Learning Algorithms Based Grid Search Algorithm for Epileptic Seizure Zone Detection

SSRN

引用

SSRN 2025年

作者： Özer, Ezgi Kinataş, Ahmet Furkan YıĞit, Emine Demir, Hamza Bedirhan Birinci, Alper Piri Reis University Engineering Faculty Computer Engineering Department İstanbul34940 Turkey İstinye University Engineering and Natural Science Faculty Software Engineering Department İstanbul34396 Turkey

Background and objective: Epilepsy is among the most prevalent illnesses of the central nervous system. This condition results in frequent, uncontrolled seizures that happen suddenly and are caused by a variety of trigger factors, including brain injury, physiological, genetic, etc. Involuntary spasms or distraction during seizures can cause severe bodily harm or even death for epileptics. In this paper, an effective method for accurately classifying Electroencephalogram (EEG) data for the early identification of epileptic seizures is ***: The suggested process essentially hybridizes several statistical data, discrete wavelet transformations (DWT), machine learning algorithms, and feature selection techniques independently. Through the use of DWT, the automated multi-resolution signal processing approach decomposes EEG signals into detail and approximation coefficients after splitting them into detailed parts with varying window sizes to guarantee an accurate classification performance. Statistical latent features are extracted from these coefficients that describe the nonlinear and dynamical patterns in the signals. Feature selection techniques were used to reduce the dimension of the feature matrix while highlighting the important elements. Different classifier structures were developed to classify input matrices. For all classifiers, the optimal hyperparameters were found using grid search techniques. Performance metrics for classification were calculated to assess the model's ***: In the analysis, to compare the proposed procedure with the other approaches in terms of detecting the epileptic behaviors correctly, a benchmark data set from the University of Bonn database was used. The results showed that the proposed approach can estimate more robust models concerning performance metrics and information criteria in classifying EEG signals. Also, the most important frequency bands were detected to distinguish EEG ***: Th

关键词： Discrete wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Constrained Over-the-Air Model Updating for Wireless Online Federated Learning with Delayed Information

arXiv

引用

arXiv 2025年

作者： Wang, Juncheng Liu, Yituo Liang, Ben Dong, Min Department of Computer Science Hong Kong Baptist University Hong Kong Department of Electrical and Computer Engineering University of Toronto Canada Department of Electrical Computer and Software Engineering Ontario Tech University Canada

We study online federated learning over a wireless network, where the central server updates an online global model sequence to minimize the time-varying loss of multiple local devices over time. The server updates the global model through over-the-air model-difference aggregation from the local devices over a noisy multiple-access fading channel. We consider the practical scenario where information on both the local loss functions and the channel states is delayed, and each local device is under a time-varying power constraint. We propose Constrained Over-the-air Model Updating with Delayed infOrmation (COMUDO), where a new lower-and-upper-bounded virtual queue is introduced to counter the delayed information and control the hard constraint violation. We show that its local model updates can be efficiently computed in closed-form expressions. Furthermore, through a new Lyapunov drift analysis, we show that COMUDO provides bounds on the dynamic regret, static regret, and hard constraint violation. Simulation results on image classification tasks under practical wireless network settings show substantial accuracy gain of COMUDO over state-of-the-art approaches, especially in the low-power region. Copyright © 2025, The Authors. All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Efficient Fingerprint Database Organization for Optimized Recognition

Efficient Fingerprint Database Organization for Optimized Re...

引用

Emerging Smart Computing and Informatics (ESCI), Conference on

作者： Jeffrey Jeeva Dhas Adhip Koshy Nishit Dua D. Ajitha Department of Software Systems School of Computer Science and Engineering Vellore Institute of Technology Vellore Tamil Nadu India Department of Software Systems School of Computer Science and Engineering Vellore Institute of Technology Vellore Vellore India

ISBN: (数字)9798331515683

ISBN: (纸本)9798331515690

A popular biometric identification method, renowned for its dependability and individuality in personal identification, is fingerprint recognition. This article presents an efficient fingerprint identification system that significantly reduces processing times for 100,000 fingerprints from 8 minutes to less than 5 seconds by utilizing multithreading, ANSI 378 templates, and Source AFIS. The system overcomes the time-consuming image-to-template conversion procedure and increases overall storage efficiency by utilizing parallel processing and preformatted templates (ANSI/ISO). The fingerprint matching system incorporates Directional Field, CNN models for identifying fingerprint types, and Finger Code methods to classify fingerprints, improving precision and efficiency.

关键词： Technological innovation Multithreading Law enforcement System performance Organizations Machine learning Fingerprint recognition Biometric identification Real-time systems Informatics

来源：评论

学校读者我要写书评

暂无评论

A Transfer Learning-Based Framework for Enhanced Classification of Perceived Mental Stress Using EEG Spectrograms

引用

IEEE Access 2025年 13卷 89266-89286页

作者： Sheharyar Khan Sadam Hussain Noorani Jaroslav Frnda Usman Rauf Aamir Arsalan Sanay Muhammad Umar Saeed Department of Computer Engineering University of Engineering and Technology Taxila Pakistan Department of Computer Science HITEC University Taxila Pakistan Department of Quantitative Methods and Economic Informatics Faculty of Operation and Economics of Transport and Communications University of Žilina Žilina Slovakia Department of Software Engineering Fatima Jinnah Women University Rawalpindi Pakistan

Stress is a significant health concern, impacting both physical and mental well-being. Prolonged exposure to stress can lead to numerous physical health issues, including cardiovascular diseases, and a weakened immune system. This study presents a novel methodology for classifying perceived mental stress using electroencephalography (EEG) signals. By utilizing the publicly available Leipzig Study for Mind-Body-Emotion Interactions dataset, we analyze EEG data collected from 53 participants over a 7-minute resting-state duration. Our approach involves transforming EEG signals into spectrograms using the Short-Time Fourier Transform (STFT), resulting in a time-frequency representation of the input signals. We employ transfer learning to fine-tune three pre-trained deep neural networks i.e., ResNet50, EfficientNetB0, and DenseNet121 for classifying stress into two and three levels. Our findings demonstrate that the ResNet50 model achieves superior classification accuracies of 95.80% and 86.02% for two and three-level stress classification, respectively, outperforming existing state-of-the-art methods. This study is the first to utilize STFT-generated spectrograms and transfer learning for perceived stress classification, highlighting the efficacy of deep learning techniques in quantifying perceived mental stress through non-invasive EEG recordings. Our results indicate that the proposed method can significantly enhance the accuracy of stress classification frameworks, offering potential improvements in mental health assessment and intervention strategies.

关键词： Human factors Electroencephalography Accuracy Stress measurement Spectrogram Physiology Transfer learning Feature extraction Brain modeling Electrodes

来源：评论

学校读者我要写书评

暂无评论

The Myth of Immutability: A Multivocal Review on Smart Contract Upgradeability

arXiv

引用

arXiv 2025年

作者： Qasse, Ilham Ali, Isra M. Ahmed, Nafisa Hamdaqa, Mohammad Jónsson, Björn Þór Department of Computer Science Reykjavik University Reykjavik Iceland College of Science and Engineering Hamad Bin Khalifa University Doha Qatar Department of Computer Engineering and Software Engineering Polytechnique Montreal Montreal Canada

Background: The immutability of smart contracts on blockchain platforms like Ethereum promotes security and trustworthiness but presents challenges for updates, bug fixes, or adding new features post-deployment. These limitations can lead to vulnerabilities and outdated functionality, impeding the evolution and maintenance of decentralized applications. Despite various upgrade mechanisms proposed in academic research and industry, a comprehensive analysis of their trade-offs and practical implications is lacking. Aims: This study aims to systematically identify, classify, and evaluate existing smart contract upgrade mechanisms, bridging the gap between theoretical concepts and practical implementations. It introduces standardized terminology and evaluates the trade-offs of different approaches using software quality attributes. Methods: We conducted a Multivocal Literature Review (MLR) to analyze upgrade mechanisms from both academic research and industry practice. We first establish a unified definition of smart contract upgradeability and identify core components essential for understanding the upgrade process. Based on this definition, we classify existing methods into full upgrade and partial upgrade approaches, introducing standardized terminology to harmonize the diverse terms used in the literature. We then characterize each approach and assess its benefits and limitations using software quality attributes such as complexity, flexibility, security, and usability. Results: The analysis highlights significant trade-offs among upgrade mechanisms, providing valuable insights into the benefits and limitations of each approach. These findings guide developers and researchers in selecting mechanisms tailored to specific project requirements. Conclusions: By offering a comprehensive evaluation that bridges theoretical concepts and practical implementations, this study contributes to the development of more resilient, adaptable, and secure decentralized systems. © 202

关键词： software quality

来源：评论

学校读者我要写书评

暂无评论

ViT Enhanced Privacy-Preserving Secure Medical Data Sharing and Classification

ViT Enhanced Privacy-Preserving Secure Medical Data Sharing ...

引用

Consumer Communications and Networking Conference, CCNC IEEE

作者： Al Amin Kamrul Hasan Sharif Ullah M. Shamim Hossain Department of Electrical & Computer Engineering College of Engineering Tennessee State University TN USA Department of Computer Science & Engineering College of Science and Engineering University of Central Arkansas Conway AR USA Department of Software Engineering College of Computer and Information Sciences King Saud University Riyadh Saudi Arabia

ISBN: (数字)9798331508050

ISBN: (纸本)9798331508067

Privacy-preserving and secure data sharing are critical for medical image analysis while maintaining accuracy and minimizing computational overhead are also crucial. Applying existing deep neural networks (DNNs) to encrypted medical data is not always easy and often compromises performance and security. To address these limitations, this research introduces a secure framework consisting of a learnable encryption method based on block-pixel operation to encrypt the data and subsequently integrate it with the Vision Transformer (ViT). The proposed framework ensures data privacy and security by creating unique scrambling patterns per key, providing robust performance against leading bit attacks and minimum difference attacks.

关键词： Data privacy computer vision Image analysis Accuracy Artificial neural networks Transformers Robustness Encryption Biomedical imaging

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：