检索结果-内蒙古大学图书馆

xCViT:Improved Vision Transformer Network with Fusion of CNN and Xception for Skin Disease Recognition with Explainable AI

引用

computers, Materials & Continua 2025年第4期83卷 1367-1398页

作者： Armughan Ali Hooria Shahbaz Robertas Damaševicius Department of Electrical Engineering Wah Engineering CollegeUniversity of WahWah Cantt47040Pakistan Department of Computer Science HITEC UniversityTaxila47080Pakistan Department of Applied Informatics Vytautas Magnus UniversityKaunas44309Lithuania

Skin cancer is the most prevalent cancer globally,primarily due to extensive exposure to Ultraviolet(UV)*** identification of skin cancer enhances the likelihood of effective treatment,as delays may lead to severe tumor *** study proposes a novel hybrid deep learning strategy to address the complex issue of skin cancer diagnosis,with an architecture that integrates a Vision Transformer,a bespoke convolutional neural network(CNN),and an Xception *** were evaluated using two benchmark datasets,HAM10000 and Skin Cancer *** the HAM10000,the model achieves a precision of 95.46%,an accuracy of 96.74%,a recall of 96.27%,specificity of 96.00%and an F1-Score of 95.86%.It obtains an accuracy of 93.19%,a precision of 93.25%,a recall of 92.80%,a specificity of 92.89%and an F1-Score of 93.19%on the Skin Cancer ISIC *** findings demonstrate that the model that was proposed is robust and trustworthy when it comes to the classification of skin *** addition,the utilization of Explainable AI techniques,such as Grad-CAM visualizations,assists in highlighting the most significant lesion areas that have an impact on the decisions that are made by the model.

关键词： Skin lesions vision transformer CNN Xception deep learning network fusion explainable AI Grad-CAM skin cancer detection

来源：评论

学校读者我要写书评

暂无评论

Intelligent Omni-Surface-Aided Integrated Sensing and Communications Based on Deep Reinforcement Learning with Knowledge Transfer

引用

IEEE Transactions on Wireless Communications 2025年第5期24卷 4344-4360页

作者： Ye, Xiaowen Mao, Yuyi Yu, Xianghao Fu, Liqun City University of Hong Kong Department of Electrical Engineering Hong Kong Hong Kong Macau University of Science and Technology School of Computer Science and Engineering Taipa China Xiamen University School of Informatics Xiamen China

This paper investigates an intelligent omni-surface (IOS)-assisted integrated sensing and communication (ISAC) system, where a base station provides both target sensing and communication services with an IOS. The sensing signal-to-noise ratio (SNR) is maximized while satisfying the communication requirement by optimizing IOS configurations. Conventional approaches typically need real-time and accurate channel state information (CSI) and have high computational complexity, making them difficult to implement in realistic systems. To circumvent this problem, this paper puts forth a new framework based on deep reinforcement learning (DRL) with knowledge transfer. In particular, an online learning scheme called Deep reinforcement learning IOS-ISAC (DeepOSC), is first proposed to optimize the reflecting and refracting coefficients of the IOS. Thereafter, to enable powerful reasoning and fast decision-making, we incorporate an echo state network (ESN) with separate output into DeepOSC. To further accelerate convergence, two transfer learning approaches, namely staged policy reuse (SPR) and staged policy distillation (SPD), are developed to guide the learning process of a newly deployed agent by leveraging policies of pre-trained agents. Numerical results show that compared to various benchmarks, DeepOSC attains significant sensing and communication performance gains and is more robust against outdated CSI coefficients. In addition, in comparison to conventional neural networks, ESN shortens the run-time of DeepOSC by more than ten times and is more efficient for temporal inference. Besides, we demonstrate the capabilities of SPR and SPD in accelerating the convergence of DeepOSC. 2002-2012 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Large-scale Multimodal Study for Predicting Mortality Risk Using Minimal and Low Parameter Models and Separable Risk Assessment

引用

IEEE Journal of Biomedical and Health informatics 2025年第5期29卷 3762-3771页

作者： Cerna, Alvaro E. Ulloa vanMaanen, David P. Jing, Linyuan Stough, Joshua V. Patel, Aalpen A. Haggerty, Christopher M. Fornwalt, Brandon K. Pattichis, Marios S. Department of Translational Data Science and Informatics Geisinger PA17822 United States University of New Mexico Department of Electrical and Computer Engineering NM87131 United States

The majority of biomedical studies use limited datasets that may not generalize over large heterogeneous datasets that have been collected over several decades. The current paper develops and validates several multimodal models that can predict 1-year mortality based on a massive clinical dataset. Our focus on predicting 1-year mortality can provide a sense of urgency to the patients. Using the largest dataset of its kind, the paper considers the development and validation of multimodal models based on 25,137,015 videos associated with 699,822 echocardiography studies from 316,125 patients, and 2,922,990 8-lead electrocardiogram (ECG) traces from 631,353 patients. Our models allow us to assess the contribution of individual factors and modalities to the overall risk. Our approach allows us to develop extremely low-parameter models that use optimized feature selection based on feature importance. Based on available clinical information, we construct a family of models that are made available in the DISIML package. Overall, performance ranges from an AUC of 0.72 with just ten parameters to an AUC of 0.89 with under 105k for the full multimodal model. The proposed approach represents a modular neural network framework that can provide insights into global risk trends and guide therapies for reducing mortality risk. © 2013 IEEE.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

Addressing Imbalance in Health Datasets: A New Method NR-Clustering SMOTE and Distance Metric Modification

引用

computers, Materials & Continua 2025年第2期82卷 2931-2949页

作者： Hairani Hairani Triyanna Widiyaningtyas Didik Dwi Prasetya Afrig Aminuddin Department of Electrical Engineering and Informatics Faculty of EngineeringUniversitas Negeri MalangMalang65145Indonesia Department of Computer Science Universitas BumigoraMataram83127Indonesia Department of Computer Graphic and Multimedia Faculty of ComputingCollege of Computing and Applied SciencesUniversiti Malaysia Pahang Al-Sultan AbdullahPekan26600Malaysia

An imbalanced dataset often challenges machine learning, particularly classification methods. Underrepresented minority classes can result in biased and inaccurate models. The Synthetic Minority Over-Sampling Technique (SMOTE) was developed to address the problem of imbalanced data. Over time, several weaknesses of the SMOTE method have been identified in generating synthetic minority class data, such as overlapping, noise, and small disjuncts. However, these studies generally focus on only one of SMOTE’s weaknesses: noise or overlapping. Therefore, this study addresses both issues simultaneously by tackling noise and overlapping in SMOTE-generated data. This study proposes a combined approach of filtering, clustering, and distance modification to reduce noise and overlapping produced by SMOTE. Filtering removes minority class data (noise) located in majority class regions, with the k-nn method applied for filtering. The use of Noise Reduction (NR), which removes data that is considered noise before applying SMOTE, has a positive impact in overcoming data imbalance. Clustering establishes decision boundaries by partitioning data into clusters, allowing SMOTE with modified distance metrics to generate minority class data within each cluster. This SMOTE clustering and distance modification approach aims to minimize overlap in synthetic minority data that could introduce noise. The proposed method is called “NR-Clustering SMOTE,” which has several stages in balancing data: (1) filtering by removing minority classes close to majority classes (data noise) using the k-nn method;(2) clustering data using K-means aims to establish decision boundaries by partitioning data into several clusters;(3) applying SMOTE oversampling with Manhattan distance within each cluster. Test results indicate that the proposed NR-Clustering SMOTE method achieves the best performance across all evaluation metrics for classification methods such as Random Forest, SVM, and Naїve Bayes, compared t

关键词： SMOTE modification Clustering-SMOTE manhattan distance

来源：评论

学校读者我要写书评

暂无评论

Feature analysis and selection for BGP anomaly detection 28

Feature analysis and selection for BGP anomaly detection

引用

28th Conference on Innovation in Clouds, Internet and Networks, ICIN 2025

作者： Hakimi, Rifqy Reed, Martin J. School of Computer Science and Electronic Engineering University of Essex Colchester United Kingdom School of Electrical Engineering and Informatics Bandung Institute of Technology Bandung Indonesia

ISBN: (纸本)9798331542399

Automated incidence reporting requires accurate and timely anomaly detection. This paper considers this in the context of the Border Gateway Protocol (BGP). BGP is crucial for Internet routing but is vulnerable to attacks due to a lack of widespread authentication. While rare, BGP disruptions can be highly detrimental to Internet performance and security. Detecting BGP anomalies helps network operators protect their networks and improve Internet reliability. In this work, we investigate and develop anomaly detection for BGP using Machine-Learning. We extract relevant features of BGP control plane messages from RIPE RIS and RouteViews public datasets. After extracting features, we employ various feature selection algorithms to extract the most relevant features and explore balancing the datasets. Finally, we explore detection latency, an important operating parameter for automated anomaly detection. The results show that BGP anomalies can be detected in the order of seven minutes when using real attack data and that the observation point for the BGP data has a significant effect on anomaly detection. © 2025 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Attention-Driven Emotion Recognition in EEG: A Transformer-Based Approach With Cross-Dataset Fine-Tuning

引用

IEEE Access 2025年 13卷 69369-69394页

作者： Ghous, Ghulam Najam, Shaheryar Alshehri, Mohammed Alshahrani, Abdulmonem AlQahtani, Yahya Jalal, Ahmad Liu, Hui Riphah International University Department of Electrical and Computer Engineering Islamabad44000 Pakistan Bahria University Department of Electrical Engineering Islamabad44000 Pakistan King Khalid University Department of Computer Science Abha61421 Saudi Arabia King Khalid University Department of Informatics and Computer Systems Abha61421 Saudi Arabia Air University Faculty of Computer Science and AI Islamabad44000 Pakistan Korea University College of Informatics Department of Computer Science and Engineering Seoul02841 Korea Republic of University of Bremen Cognitive Systems Laboratory Bremen28359 Germany Guodian Nanjing Automation Company Ltd. Nanjing210032 China

Emotion recognition from EEG (electroencephalogram) signals is crucial in mental health diagnostics and human-computer interaction but is often hindered by high dimensionality, noise, and complex temporal dependencies in the data. This paper presents a novel approach that integrates transformer models, attention mechanisms, and transfer learning to enhance emotion recognition accuracy from EEG signals. The proposed methodology consists of two phases: Attention Enhanced Base Model Development (AE-BMD) and Cross-Dataset Fine Tuning Adaptation (CD-FTA). In the AE-BMD phase, the base model is developed and trained on the SEED-IV dataset (15 participants, 62 EEG channels), achieving an accuracy of 84%, with an average precision of 84.75%, recall of 84% and F1-score of 84%. This phase employs optimized feature extraction from key EEG frequency bands (Delta, Theta, Alpha, Beta, Gamma) using techniques such as MFCC, GFCC, power spectral density, and Hjorth parameters. A transformer encoder with integrated spectral and temporal attention mechanisms captures intricate patterns and long-range dependencies within the EEG signals. In the CD-FTA phase, the model undergoes fine-tuning on the SEED-V dataset (20 participants, 62 channels) leading to an improved accuracy of 90%, with an average precision of 90.6%, recall of 90.6%, and F1-score of 90.6%. The model’s generalization is further validated on the MPED dataset (23 participants, 62 channels, seven emotion classes), achieving 79%, with an average precision of 79.3%, recall of 79.3% and F1-score of 79.1% across diverse emotional states. This cross-dataset adaptation leverages transfer learning to enhance the model’s generalization across different emotional states and EEG datasets. Experimental results show that the proposed approach outperforms traditional methods, achieving superior accuracy and robustness in emotion recognition tasks. This work advances emotion recognition systems by addressing challenges in EEG signal proc

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

A Benchmark for Multi-Speaker Anonymization

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 3819-3833页

作者： Miao, Xiaoxiao Tao, Ruijie Zeng, Chang Wang, Xin Singapore Institute of Technology Singapore567739 Singapore National University of Singapore Electrical and Computer Engineering Department Queenstown 117583 Singapore National Institute of Informatics Chiyoda-ku Tokyo101-8340 Japan

Privacy-preserving voice protection approaches primarily suppress privacy-related information derived from paralinguistic attributes while preserving the linguistic content. Existing solutions focus particularly on single-speaker scenarios. However, they lack practicality for real-world applications, i.e., multi-speaker scenarios. In this paper, we present an initial attempt to provide a multi-speaker anonymization benchmark by defining the task and evaluation protocol, proposing benchmarking solutions, and discussing the privacy leakage of overlapping conversations. The proposed benchmark solutions are based on a cascaded system that integrates spectral-clustering-based speaker diarization and disentanglement-based speaker anonymization using a selection-based anonymizer. To improve utility, the benchmark solutions are further enhanced by two conversation-level speaker vector anonymization methods. The first method minimizes the differential similarity across speaker pairs in the original and anonymized conversations, which maintains original speaker relationships in the anonymized version. The other minimizes the aggregated similarity across anonymized speakers, which achieves better differentiation between speakers. Experiments conducted on both non-overlap simulated and real-world datasets demonstrate the effectiveness of the multi-speaker anonymization system with the proposed speaker ***, we analyzed overlapping speech regarding privacy leakage and provided potential solutions © 2005-2012 IEEE.

关键词： Anonymity

来源：评论

学校读者我要写书评

暂无评论

Sobol Sequence Optimization for Hardware-Efficient Vector Symbolic Architectures

引用

IEEE Transactions on computer-Aided Design of Integrated Circuits and Systems 2025年第3期44卷 937-950页

作者： Aygun, Sercan Najafi, Hassan University of Louisiana at Lafayette School of Computing and Informatics LafayetteLA70503 United States Case Western Reserve University Electrical Computer and Systems Engineering Department ClevelandOH44106 United States

Hyperdimensional computing (HDC) is an emerging computing paradigm with significant promise for efficient and robust learning. In HDC, objects are encoded with high-dimensional vector symbolic sequences called hypervectors. The quality of hypervectors, defined by their distribution and independence, directly impacts the performance of HDC systems. Despite a large body of work on the processing parts of HDC systems, little to no attention has been paid to data encoding and the quality of hypervectors. Most prior studies have generated hypervectors using inherent random functions, such as MATLAB’s or Python’s random function. This work introduces an optimization technique for generating hypervectors by employing quasi-random sequences. These sequences have recently demonstrated their effectiveness in achieving accurate and low-discrepancy data encoding in stochastic computing systems. The study outlines the optimization steps for utilizing Sobol sequences to produce high-quality hypervectors in HDC systems. An optimization algorithm is proposed to select the most suitable Sobol sequences via indexes for generating minimally correlated hypervectors, particularly in applications related to symbol-oriented architectures. The performance of the proposed technique is evaluated in comparison to two traditional approaches of generating hypervectors based on linear-feedback shift registers and MATLAB random functions. The evaluation is conducted for three applications: 1) language;2) headline;and 3) medical image classification. Our experimental results demonstrate accuracy improvements of up to 10.79%, depending on the vector size. Additionally, the proposed encoding hardware exhibits reduced energy consumption and a superior area-delay product. © 2024 IEEE.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

A Two-Pronged Approach to Security of Discrete-Event Systems

引用

Journal of Systems Science & Complexity 2025年第1期38卷 178-209页

作者： MATSUI Shoma RUDIE Karen CAI Kai Department of Electrical and Computer Engineering&the Ingenuity Labs Research Institute Queen's UniversityKingston K7L 3N6Canada Department of Core Informatics Osaka Metropolitan UniversityOsaka 545-0051Japan

This paper investigates a security problem of simultaneously addressing two types of attacks:Eavesdropping and *** authors model the target system as a discrete-event system(DES)with subsets of concealable events and protectable events,in order to make the proposed methodology applicable to various practical systems and employ two existing works of DES security:Degree of opacity and state ***,the authors consider that all protectable events are observable,and some observable events are *** addition,protectable events cannot be protected once they are *** such a constraint,the goal is to figure out which events to conceal and which transitions to protect so that the prescribed requirements of degree of opacity and state protection are *** this work the authors decide which events to conceal as all transitions of a given event label are concealed or not *** proposed problem formulation also requires a solution to only involve absolutely necessary protectable events in order for the system to avoid superfluous protection *** authors first examine a general version of our security problem with an intuitive algorithm to compute acceptable solutions,and then present a special version which results in a reduced computation time compared to the general version.

关键词： Degree of opacity discrete-event systems security state protection

来源：评论

学校读者我要写书评

暂无评论

BEDiff: denoising diffusion probabilistic models for building extraction

引用

Optoelectronics Letters 2025年第5期21卷 298-305页

作者： LEI Yanjing WANG Yuan CHAN Sixian HU Jie ZHOU Xiaolong ZHANG Hongkai College of Computer Science and Technology Zhejiang University of Technology Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering College of Computer and Information China Three Gorges University Key Laboratory of Intelligent Informatics for Safety & Emergency of Zhejiang Province Wenzhou University College of Electrical and Information Engineering Quzhou University School of Electronic and Information Engineering Anhui Jianzhu University

Accurately identifying building distribution from remote sensing images with complex background information is challenging. The emergence of diffusion models has prompted the innovative idea of employing the reverse denoising process to distill building distribution from these complex backgrounds. Building on this concept, we propose a novel framework, building extraction diffusion model(BEDiff), which meticulously refines the extraction of building footprints from remote sensing images in a stepwise fashion. Our approach begins with the design of booster guidance, a mechanism that extracts structural and semantic features from remote sensing images to serve as priors, thereby providing targeted guidance for the diffusion process. Additionally, we introduce a cross-feature fusion module(CFM) that bridges the semantic gap between different types of features, facilitating the integration of the attributes extracted by booster guidance into the diffusion process more effectively. Our proposed BEDiff marks the first application of diffusion models to the task of building extraction. Empirical evidence from extensive experiments on the Beijing building dataset demonstrates the superior performance of BEDiff, affirming its effectiveness and potential for enhancing the accuracy of building extraction in complex urban landscapes.

关键词： A

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：