检索结果-内蒙古大学图书馆

Lead-cnn: lightweight enhanced dimension reduction convolutional neural network for brain tumor classification

International Journal of Machine Learning and Cybernetics 2025年 1-20页

作者： Khan, Saif Ur Rehman Asif, Sohaib Bilal, Omair Rehman, Hafeez Ur School of Computer Science and Engineering Central South University 932 Lushan S Rd Yuelu District Hunan Changsha China School of Computing and Data Sciences Oryx Universal College with Liverpool John Moores University Doha Qatar Department of Computer Science National University of Computer and Emerging Sciences Islamabad Pakistan Taizhou Key Laboratory of Minimally Invasive Interventional Therapy & Artificial Intelligence Taizhou Campus of Zhejiang Cancer Hospital Taizhou Cancer Hospital Taizhou317502 China Chinese Academy of Sciences Taizhou317502 China Chinese Academy of Sciences Hangzhou310022 China

In many nations, brain tumors are among the most serious medical conditions affecting both adults and children. Timely and accurate diagnosis is crucial, as the life of a patient may be shortened if brain tumors are not identified promptly. When brain tumors are diagnosed accurately and promptly, patients can receive the appropriate treatment. One of the challenging aspects of dealing with brain tumors is their diversity, making it difficult to categorize them accurately due to their varying characteristics. Deep learning (DL) models have emerged as a valuable tool in classifying brain tumors into distinct groups. However, these models face challenges related to accuracy and computational cost, necessitating advancements in brain image classification. Our primary aim is to develop a lightweight convolutional neural network (CNN) architecture that offers high efficiency, we incorporate a modified dimension reduction block with CNN to detect brain cancers with minimal computational resources while maintaining a high level of accuracy. In this research, we introduce a novel lightweight network called LEAD-CNN, designed specifically for the accurate and reliable identification of brain cancers from magnetic resonance images (MRI). Our approach begins by evaluating the effectiveness of several pre-trained DL models, including ResNet-101, VGG-19, MobileNet-V1, and DenseNet-201, which have been recommended for the classification of medical images. Subsequently, we propose a dimension reduction base CNN architecture with a lightweight design aimed at achieving a low false-negative rate in tumor image classification for patients. To assess the performance of the proposed model, we utilize a benchmark brain MRI dataset (Kaggle) comprising 7023 images, encompassing three different types of brain tumors long with normal brain images. Our experimental results are demonstrating an overall classification accuracy of 98.70%, with an average recall, F1-score, and precision of 98.60%

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Exploring Invertible Encoding for Deep Video Compression

引用

IEEE Transactions on Broadcasting 2025年

作者： Guo, Haifeng Kwong, Sam Zhou, Mingliang City University of Hong Kong Department of Computer Science Hong Kong Hong Kong Lingnan University School of Data Science Hong Kong Hong Kong Chongqing University School of Computer Science Chongqing400044 China

Deep video compression methods typically use autoencoder-style networks for encoding and decoding, which can result in the loss of information during encoding that cannot be retrieved during decoding. To address this issue, recent work has explored the use of invertible neural networks for enhanced invertible encoding, which has successfully mitigated spatial information loss for better image compression. In this paper, we propose a new approach that extends invertible encoding to temporal information and introduces an encoding-decoding network for deep video compression. Our network incorporates a novel attentive channel squeeze module to improve compression performance while also leveraging a conditional coding framework for motion compression. The entire framework is optimized via a single loss function that balances bit cost and frame quality. The experimental results demonstrate the effectiveness of our approach, which achieves 15.45%/57.92% bit savings in terms of PSNR/MS-SSIM compared with the high-efficiency video coding low-delay P configuration. © 2025 IEEE.

关键词： Decoding

来源：评论

学校读者我要写书评

暂无评论

Discovering multi-agent systems for resource-centric business process simulation

引用

Process science 2025年第1期2卷 1-26页

作者： Kirchdorfer, Lukas Blümel, Robert Kampik, Timotheus van der Aa, Han Stuckenschmidt, Heiner SAP Signavio Walldorf Germany Data and Web Science Group University of Mannheim Mannheim Germany Faculty of Computer Science University of Vienna Vienna Austria Department of Computing Science Umeå University Umeå Sweden

Business process simulation (BPS) is a powerful tool for estimating process performance across different scenarios, offering critical support for organizational process redesign and optimization. Traditional BPS approaches predominantly rely on a control-flow-first perspective by enriching a process model with simulation parameters. While these approaches seem suitable for capturing centrally orchestrated processes, such as those managed by workflow systems, they fall short of accurately reflecting real-world processes characterized by decentralized decision-making and distinct resource behaviors. To overcome this limitation, we propose AgentSimulator, a resource-first BPS approach that discovers a multi-agent system from an event log. By modeling the distinct behaviors and interaction patterns of individual resources, AgentSimulator effectively simulates the underlying process. Our approach automatically identifies whether resource behavior is rather orchestrated or autonomous, adapting to the specific decision-making structure of the process. Experimental results reveal that AgentSimulator achieves state-of-the-art simulation accuracy while ensuring high adaptability to various process types.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Thalaseemia Prediction Using Deep Neural Networks: A Step Towards Personalized Medicine

Thalaseemia Prediction Using Deep Neural Networks: A Step To...

引用

Machine Learning and Autonomous Systems (ICMLAS), International Conference on

作者： V. Pream Sudha S. Karpagavalli K. Sathiyakumari A.C. Sountharraj Department of Computer Science with Data Analytics Dr.N. G.P Arts and Science College Coimbatore India Department of Computer Science PSGR Krishnammal College for Women Coimbatore India

ISBN: (数字)9798331505745

ISBN: (纸本)9798331505752

Thalaseemia is a series of genetic hemolytic diseases brought on by defective hemoglobin synthesis. It is prevalent in many Asian, African, and Mediterranean nations. Four million Indians are Thalaseemia carriers, and more than 100,000 are actively combating the condition, according to the World Health Organization. It is specifically taken under investigation because cases of Thalassemia that have clinical, hematologic characteristics cannot be effectively separated from cases of sickle cell anemia and are misdiagnosed as such. The precise profiles of the genetic and pathophysiological mechanisms of patients with Thalassemia are not yet well defined, despite the fact that they are clinically treated identically. Even among people with the same genotype, there are differences in the severity of the clinical phenotype. The frequency of complications changes over time, both within and between individuals. Patients display a striking variation in severity, from a milder clinical course to severe transfusion dependency and progressive organ destruction. This study focuses on the creation of a discriminative model for Thalaseemia classification. In this study, the prediction of Thalassemia is modeled as a pattern classification issue, and the models are constructed using deep learning approaches. The hybrid CNN-Long Short Term Memory (LSTM) based Thalaseemia prediction model has demonstrated improved performance, according to the performance evaluation of these models.

关键词： Deep learning Time-frequency analysis Phenotypes Precision medicine Predictive models Feature extraction data models Convolutional neural networks Long short term memory Blood

来源：评论

学校读者我要写书评

暂无评论

A Framework for Video Based Sign Language Interpretation Using Machine Learning and Statistical Methods

A Framework for Video Based Sign Language Interpretation Usi...

引用

Sentiment Analysis and Deep Learning (ICSADL), International Conference on

作者： Swathi D Susmetha K Sreeji. S S. Mangairkarasi B.E Computer Science and Engineering with Data Science Sathyabama Institute of Science and Technology Semmancheri Chennai TamilNadu India Department of Computer Science and Engineering Sathyabama Institute of Science and Technology Semmancheri Chennai TamilNadu India

ISBN: (数字)9798331523923

ISBN: (纸本)9798331523930

The Sign Language Recognition System has been designed to capture video input, process it to detect hand gestures, and translate these gestures into readable text. The project consists of several key components and steps: Video Processing: Using OpenCV, the system captures frames from the video input. MediaPipe processes these frames to detect and track hand landmarks in real time. OpenCV capabilities allow for efficient frame extraction and basic image processing tasks such as resizing and normalization. Hand Detection and Tracking: MediaPipe pre-trained models identify and track hand movements within the video frames. The accurate detection and tracking of the hand movements are critical for the subsequent recognition of the sign language gestures. Sign Language Recognition: The core system is the deep learning model, trained using the TensorFlow and Keras on a dataset of sign language gestures. The model learns to classify the detected hand movements into corresponding sign language characters or words. Convolutional Neural Networks (CNNs) are typically used for task due to their effectiveness in image recognition tasks. Text Display: Once the system recognizes the signs, it converts them into text and displays the output. This can be done through a console output or a graphical user interface (GUI) built with Tkinter. The GUI provides a user friendly experience, allowing users to see the translated text in real-time.

关键词： Hands Deep learning Sign language Translation Tracking Text recognition Speech recognition Streaming media Real-time systems Graphical user interfaces

来源：评论

学校读者我要写书评

暂无评论

Multi-Agent Systems for Autonomous IoT Network Management Using Distributed Reinforcement Learning 3

Multi-Agent Systems for Autonomous IoT Network Management Us...

引用

3rd International Conference on Intelligent Systems, Advanced Computing, and Communication, ISACC 2025

作者： Neelamegam, G. Venkatesan, Rajaram Ramya, S.R. Ramya, R.S. Akshya, J. Sundarrajan, M. Choudhry, Mani Deepak VSB College of Engineering and Technology Department of Computer Science and Engineering Tamil Nadu Coimbatore India Builders Engineering College Department of Artificial Intelligence and Data Science Tamil Nadu Tirupur India Dr NGP Institute of Technology Department of Computer Science and Engineering Tamil Nadu Coimbatore India Sri Ramakrishna Engineering College Department of Computer Science and Engineering Tamil Nadu Coimbatore India SRM Institute of Science and Technology Department of Computational Intelligence Tamil Nadu Chennai India SRM Institute of Science and Technology Department of Networking and Communications Tamil Nadu Chennai India SRM Institute of Science and Technology Department of Computing Technologies Tamil Nadu Chennai India

ISBN: (纸本)9798331523893

The ever-growing complexity of IoT networks ignited by their wide scale adoption in applications such as smart cities, the industrial automation, and health care, compelled to develop sophisticated yet resource efficient and scalable network management solution. Centralized systems are traditional in architecture, which leads to higher latency and bottlenecks, while decentralized systems can distribute workloads however the lack of adaptability to real time, along with the difficulty of extracting maximum resource usage, are limitations. In recent years, reinforcement learning (RL) based approaches have taken shape as a promising alternative to facilitate the automated learning and optimisation of task management. However, there are two challenges for existing RL based methods: slow convergence, inefficient priority assignment of tasks, and poor coordination amongst agents in the case of dynamic, large scale IoT environment. To solve these challenges, in this research we propose a distributed reinforcement learning based multi agent framework for managing IoT network. The framework also integrates a task prioritization mechanism to dynamically choose an optimal task allocation solution. Further validation was performed in a simulated IoT environment where real world inspired datasets were used. The proposed framework demonstrated 30% more task throughput than centralized, 25% less latency than decentralized, and 40% more energy efficiency than current RL based systems. The system was further validated using advanced visualizations such as anomaly detection maps and resource allocation efficiency graphs, demonstrating its capacity to manage dynamic load of tasks and resource allocation. The proposed framework addresses the critical limitations of current methods, through scalable, adaptive, and energy efficient framework for managing IoT network, and accordingly helps advancing the intelligent and autonomous IoT systems. © 2025 IEEE.

关键词： Resource allocation

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Strategies for Multiclass Skin Disease Classification 5th

Deep Learning Strategies for Multiclass Skin Disease Classi...

引用

5th EAI International Conference on Cognitive Computing and Cyber Physical Systems, IC4S 2024

作者： Venkat, Nakka Lokeswara Satya Tejasri, Masina Naga Vijaya Uppalapati, Padma Jyothi Rao, V. V. R. Maheswara Sripada, V. S. S. Lakshmi Murty, P. Sita Rama Department of Artificial Intelligence and Data Science Vishnu Institute of Technology Andhra Pradesh Bhimavaram534201 India Department of Computer Science and Engineering Vishnu Institute of Technology Andhra Pradesh Bhimavaram534201 India Department of CSE Shri Vishnu Engineering College for Women Andhra Pradesh Bhimavaram534201 India Andhra Pradesh Bhimavaram534201 India

ISBN: (纸本)9783031770746

Skin diseases represent a significant global health challenge, necessitating accurate and timely diagnosis for effective treatment. In this paper, we delve into the intricate realm of skin disease classification, leveraging advanced deep learning techniques. Skin diseases, ranging from common conditions like eczema and atopic dermatitis to potentially life-threatening melanoma, pose complex diagnostic challenges due to their varied visual characteristics. Accurate classification demands the ability to discern subtle patterns, textures, and features within medical images. Our paper introduces a novel methodology that includes the integration of our proposed model into the existing deep learning frameworks. Unlike conventional approaches involving the replacement of output layers, our model maintains the integrity of pre-trained structures. It achieves this by utilizing a blend of VGG16 and EfficientNet B2, incorporating dense activations for improved regularization. This novel approach seeks to improve the model’s flexibility to various skin disease patterns while maintaining accurate classification without changing the pre-trained models’ fundamental design. The model’s performance is evaluated in-depth using a variety of criteria, including accuracy, precision, recall, and F1 score. The outcomes of this analysis unveil distinct F1 scores for each disease class across different models, thereby illuminating the models’ strengths and limitations in the context of disease classification. © ICST Institute for computer sciences, Social Informatics and Telecommunications Engineering 2025.

关键词： Dermatitis

来源：评论

学校读者我要写书评

暂无评论

Feature Selection in Socio-Economic Analysis: A Multi-Method Approach for Accurate Predictive Outcomes

引用

International Journal of Crowd science 2025年第1期9卷 64-78页

作者： Al-Qerem, Ahmad Ali, Ali Mohd Jebreen, Issam Nabot, Ahmad Rajab, Mohammed Alauthman, Mohammad Aldweesh, Amjad Aburub, Faisal Alangari, Someah Alzgol, Musab Computer Science Department Faculty of Information Technology Zarqa University Zarqa13110 Jordan Communications and Computer Engineering Department Faculty of Engineering Al-Ahliyya Amman University Amman19328 Jordan University Headquarter University of Anbar Ramadi31001 Iraq Department of Information Security University of Petra Amman11196 Jordan College of Computing and Information Technology Shaqra University Shaqra11911 Saudi Arabia Department of Business Intelligence and Data Analytics University of Petra Amman11196 Jordan Computer Information Systems Department Faculty of Information Technology Isra University Amman11622 Jordan

Feature selection is a cornerstone in advancing the accuracy and efficiency of predictive models, particularly in nuanced domains like socio-economic analysis. This study explores nine distinct feature selection methods, utilizing a heart disease dataset as a representative model for complex socio-economic systems. Our findings identified four universally recognized features as critical across all selection methods. However, the divergence in significance attributed to other features by different methods underscores the inherent variability in selection techniques. When the top four features were incorporated into twelve classification models, a noticeable surge in predictive accuracy was observed, emphasizing their foundational role in enhancing model outcomes. The variations among methods stress the need for a methodical and discerning approach to feature selection, especially in data-rich socio-economic landscapes. As we venture further into an era defined by data-driven decision-making, rigour and precision in feature selection become indispensable. Future research should extend this approach to broader datasets, ensuring the robustness and adaptability of our findings. © The author(s) 2025.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Identification of Tamil Sign Language using Hamiltonian deep neural network

引用

International Journal of Systems, Control and Communications 2025年第1期16卷 62-78页

作者： Lingeswari, Ponnusamy Pavunraj, Daniel Kanagavalli, Mariappan Manikandan, Karuppasamy Department of Electronics and Communication Engineering P.S.R. Engineering College Tamil Nadu Sivakasi India Department of Artificial Intelligence and Data Science Vel Tech Multi Tech Dr. Rangarajan Dr. Sakunthala Engineering College Tamil Nadu Chennai India Department of Electronics and Communication Engineering P.S.R.R. College of Engineering Tamil Nadu Sivakasi India Department of Computer Science and Engineering P.S.R. Engineering College Tamil Nadu Sivakasi India

Sign language consisting of various hand patterns, serves as a vital communication medium for conveying messages, sharing knowledge, and expressing ideas among the deaf community. This research addresses the challenge of recognising Tamil Sign Language (TSL) by proposing a novel identification method using a Hamiltonian deep neural network (HDNN). The primary objective is to maximise the accuracy and reliability of TSL recognition, especially for real-time applications. The dataset comprises real-time images of 12 vowels, 1 Aayutha Ezhuthu, and 18 consonants, collected from 110 different signers. To address noise in the images, an iterative guided filtering (IGF) method is employed for preprocessing. Subsequently, feature extraction is performed using residual exemplars local binary pattern (RELBP). HDNN is then applied to classify the signs. The performance of the proposed approach achieves 23.32%, 25.07%, 21%, 27.53%, and 30% higher accuracy and 14.09%, 19.63%, 28%, 18.45%, and 10.54% lower error rates compared to existing methods. Copyright © 2025 Inderscience Enterprises Ltd.

关键词： Local binary pattern

来源：评论

学校读者我要写书评

暂无评论

An Optimized Multi-Objective Task Scheduling Approach for IoT Systems in the Edge-Cloud Continuum

An Optimized Multi-Objective Task Scheduling Approach for Io...

引用

Computational Intelligence Approaches and Applications (ICCIAA), International Conference on

作者： Hussam Fakhouri Fahed Alkhabbas Sadi Alawadi Feras M. Awaysheh Majed Ayyad Data Science and Artificial Intelligence Department University of Petra Amman Jordan Department of Computer Science and Media Technology Sustainable Digitalisation Research Centre Malmö University Sweden School of Computer Science Blekinge Institute of Technology Blekinge Sweden Department of Computer Science ADSLab Umeå University Umeå Sweden Department of Computer Science Birzeit University Palestine

ISBN: (数字)9798331523657

ISBN: (纸本)9798331523664

The Internet of Things (IoT) and Artificial Intelligence (AI) has enabled the development of innovative applications. The deployment of those applications is a complex process that should take into consideration multiple factors, including the applications' scale, complexity, distribution, and non-functional requirements (e.g., energy consumption, performance, and security). Moreover, deployment environments over the edge-cloud continuum are heterogeneous w.r.t. their processing capabilities, communication latencies, and energy consumption. Towards enabling efficient scheduling of tasks in such environments, we formulate the task scheduling problem as a multi-objective optimization task balancing energy efficiency and deadline adherence. To tackle this problem, we employ the Equilibrium Optimizer (EO)-a physics-inspired meta-heuristic algorithm that utilizes an equilibrium pool of top-performing solutions to guide its population toward high-quality schedules. To validate the feasibility of our approach, we run experiments where we compare our proposed approach against the multiple existing optimizers. The results demonstrate that EO exhibits a superior performance reflecting its potential to improve IoT systems' quality of service and reduce their operational costs in large-scale and time-sensitive IoT scenarios.

关键词： Energy consumption Schedules Scalability Metaheuristics Quality of service Energy efficiency Security Internet of Things Resource management Artificial intelligence

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：