检索结果-内蒙古大学图书馆

An Investigation of Frequency-Domain Pruning Algorithms for Accelerating Human Activity Recognition Tasks Based on Sensor Data

引用

computers, Materials & Continua 2024年第11期81卷 2219-2242页

作者： Jian Su Haijian Shao Xing Deng Yingtao Jiang School of Computer Jiangsu University of Science and TechnologyZhenjiang212003China Department of Electrical and Computer Engineering University of NevadaLas VegasNV 89154USA

The rapidly advancing Convolutional Neural Networks(CNNs)have brought about a paradigm shift in various computer vision tasks,while also garnering increasing interest and application in sensor-based Human Activity Recognition(HAR)***,the significant computational demands and memory requirements hinder the practical deployment of deep networks in resource-constrained *** paper introduces a novel network pruning method based on the energy spectral density of data in the frequency domain,which reduces the model’s depth and accelerates activity *** traditional pruning methods that focus on the spatial domain and the importance of filters,this method converts sensor data,such as HAR data,to the frequency domain for *** emphasizes the low-frequency components by calculating their energy spectral density ***,filters that meet the predefined thresholds are retained,and redundant filters are removed,leading to a significant reduction in model size without compromising performance or incurring additional computational ***,the proposed algorithm’s effectiveness is empirically validated on a standard five-layer CNNs backbone *** computational feasibility and data sensitivity of the proposed scheme are thoroughly ***,the classification accuracy on three benchmark HAR datasets UCI-HAR,WISDM,and PAMAP2 reaches 96.20%,98.40%,and 92.38%,***,our strategy achieves a reduction in Floating Point Operations(FLOPs)by 90.73%,93.70%,and 90.74%,respectively,along with a corresponding decrease in memory consumption by 90.53%,93.43%,and 90.05%.

关键词： Convolutional neural networks human activity recognition network pruning frequency-domain transformation

来源：评论

学校读者我要写书评

暂无评论

Deep learning based predicting urban traffic congestion with RGB-coded images using GRU-CNN and LSTM

引用

Multimedia Tools and Applications 2024年第38期83卷 86261-86280页

作者： P, Rajesh Azhagiri, M. Department of Computer Science and Engineering Vel Tech Rangarajan Dr. Sagunthala R & ampD Institute of Science and Technology Tamil Nadu Chennai India Department of Computer Science and Engineering SRM Institute of Science and Technology Ramapuram Campus Tamil Nadu Chennai600 089 India

Road traffic management requires the ability to foresee geographical congestion conditions in an urban road traffic network. The proposed investigation is aimed to envisage the presence of blockage in a specific region of geographical location using gated recurrent unit (GRU) combined with the convolutional neural network (CNN) and long short-term memory (LSTM). The data related to road traffic is extracted from the time—frequency domain of a specific urban road. The collected data is then compressed into a RGB coded image which is the input to the GRU which determines the parameters of traffic blockage in a specific region of the road. The road traffic network features are extracted from those parameters and identified using a CNN's feature extraction technology. We used LSTM model to assess the time series temporal images of the road traffic blockage and to utilize the time division to pool the flow of traffic. For model validation, CTT (Chicago Traffic Tracker) dataset network is chosen. The parameters assessed had ensured highest prediction accuracy of 98.87%, F1-score of 0.98, 0.971 of specificity, 0.987 of recall and 0.978 of precision to forecast the presence road traffic. The proposed model demonstrates a significant advancement in road traffic detection with an accuracy of 98.87%, surpassing existing models that achieve the highest accuracy of 94.3%. This represents a notable improvement of approximately 4.77% over the highest accuracy achieved by the previous models. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Securing FANET using federated learning through homomorphic matrix factorization

引用

International Journal of Information technology (Singapore) 2025年第1期17卷 17-36页

作者： Banerjee, Aiswaryya Mahato, Ganesh Kumar Chakraborty, Swarnendu Kumar Department of Computer Science and Engineering National Institute of Technology Arunachal Pradesh Jote 791113 India

As Flying Ad Hoc Networks (FANETs) continue to advance, ensuring robust security, privacy, and data reliability remains a significant challenge. This research presents a novel framework known as HE-FSMF-short for Homomorphic Encrypted Federated Secure Matrix Factorization-specifically designed to tackle these challenges. HE-FSMF integrates matrix factorization with federated learning and homomorphic encryption to enhance both security and efficiency in FANET environments. Matrix factorization, commonly used in recommendation systems, is adapted here to address the unique complexities of FANETs. By leveraging detailed feature extraction through the VGG-16 model, HE-FSMF ensures precise and secure data processing even in dynamic and high-mobility settings. The incorporation of homomorphic encryption protects data throughout cloud-based computations, maintaining privacy and integrity without compromising performance. Additionally, HE-FSMF features mechanisms to verify the accuracy and authenticity of results, which is crucial for establishing trust in distributed systems. This framework not only enhances learning efficiency and improves data transmission rates but also provides strong safeguards for sensitive information. HE-FSMF offers a robust solution for advancing FANET capabilities, making it a valuable tool for secure and efficient communication in the increasingly interconnected and rapidly evolving landscape of networked systems. © Bharati Vidyapeeth's institute of computer Applications and Management 2024.

关键词： FANET Federated learning Homomorphic encryption UAV

来源：评论

学校读者我要写书评

暂无评论

Street Light Monitoring and Controlling System Using RaspberryPi and Zigbee Protocol 7

Street Light Monitoring and Controlling System Using Raspber...

引用

2023 7th International Conference On Computing, Communication, Control And Automation, ICCUBEA 2023

作者： Gokhale, Pratik Deshpande, Shantanu Page, Saurabh Bagul, Mandar Bundele, Ashwini Pune Institute of Computer Technology Department of Computer Engineering Pune India

ISBN: (纸本)9798350304268

Street Lighting System is a crucial part of society's amenities and environment. Today's systems consume enormous amounts of electrical energy, for the automation of switches on/off of street lights. This results in a huge loss of energy and money at the same time which needs an efficient system to be developed. We propose a Smart Lighting System which replaces the traditional timer and photocell-based street lighting and outdoor lighting control. It is a complete integrated solution consisting of a Zigbee-based Node integrated with every street light fixture or outdoor lighting fixture, and operating software that uses a Master-Slave concept to control every Node (Slave) from the coordinator module (Master). With energy-saving initiatives, putting the latest technologies in use, this system has the potential to save energy as well as be an effective adaptable way of constructing a lighting system. The system will be deployed on a cloud platform through which the application is remotely accessible. © 2023 IEEE.

关键词： IOT based system Lighting System Raspberry Pi Zigbee

来源：评论

学校读者我要写书评

暂无评论

Generative Adversarial Networks for Motor Imagery Classification using Wavelet Packet Decomposition and Complex Morlet Transform

引用

Multimedia Tools and Applications 2025年 1-24页

作者： Shirodkar, Vaishali Edla, Damodar Reddy Kumari, Annu Department of Computer Science and Engineering National Institute of Technology Goa Kottamoll Plateau Goa Cuncolim403703 India Information Technology Department Goa College of Engineering Farmagudi Goa Ponda403401 India

The area of brain-computer interface research is widely spreading as it has a diverse array of potential applications. Motor imagery classification is a boon to several people with motor impairment. Low accuracy and datasets with few trial recordings present challenges for this classification. This research offers a novel approach to these problems in electroencephalography signal classification for efficient control of neuroprosthetic devices. By using wavelet packet decomposition the proposed method divides the EEG signal into eight frequency bands. Each of these bands were then evaluated by a one-dimensional convolution neural network to find the most discriminative band. Only EEG signals filtered with these bands were used to generate time-frequency images called scalograms using complex Morlet transform. These scalograms were then classified using a two-dimensional convolution neural network optimized for spatial-temporal feature extraction and classification. To mitigate data scarcity, generative adversarial networks were used for effective data augmentation. Experimental evaluation on a benchmark dataset demonstrated substantial improvement: the two-dimensional convolution neural network achieved 93.63% accuracy, surpassing the 85% achieved by the one-dimensional convolution neural network baseline. Precision, recall, and F1-score metrics validated robust performance across motor imagery tasks. Comparative analysis with the latest methods underscored the proposed model’s superiority in accuracy and reliability. Using wavelet transformations, convolution neural network architectures, and generative adversarial network-driven data augmentation presented a promising framework for advancing brain-computer interface applications, particularly in enhancing neuroprosthetic device control. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.

关键词： Wavelet decomposition

来源：评论

学校读者我要写书评

暂无评论

Adversarial-Learning-Based Taguchi Convolutional Fuzzy Neural Classifier for Images of Lung Cancer

引用

IEEE Access 2024年 12卷 72766-72776页

作者： Lin, Cheng-Jian Lin, Xue-Qian Jhang, Jyun-Yu National Chin-Yi University of Technology Department of Computer Science and Information Engineering Taichung41170 Taiwan National Taichung University of Science and Technology Department of Computer Science and Information Engineering Taichung40401 Taiwan

Deep learning technology has extensive application in the classification and recognition of medical images. However, several challenges persist in such application, such as the need for acquiring large-scale labeled data, configuring network parameters, and handling excessive network parameters. To address these challenges, in this study, we developed an adversarial-learning-based Taguchi convolutional fuzzy neural classifier (AL-TCFNC) for classifying malignant and benign lung tumors displayed in computed tomography images. In the framework of the developed AL-TCFNC, a fuzzy neural classifier replaces a conventional fully connected network, thereby reducing the number of network parameters and the training duration. To reduce experimental cost and training time, the Taguchi method was used. This method helps to identify the optimal combination of model parameters through a small number of experiments. The transfer learning of models across databases often results in subpar performance because of the paucity of labeled samples. To resolve this problem, we used a combination of maximum mean discrepancy and cross-entropy for adversarial learning with the proposed model. Two data sets, namely the SPIE-AAPM Lung CT Challenge data set and LIDC-IDRI Lung Imaging Research data set, were used to validate the AL-TCFNC model. When the AL-TCFNC model was used for transfer learning, it exhibited an accuracy rate of 89.55% and outperformed other deep learning models in terms of classification performance. © 2013 IEEE.

关键词： Fuzzy neural networks

来源：评论

学校读者我要写书评

暂无评论

Attention Contours: A Comparative Study on 2D and Virtual Reality Attention Profile Assessment Tests 5

Attention Contours: A Comparative Study on 2D and Virtual Re...

引用

5th International Conference on Data Intelligence and Cognitive Informatics, ICDICI 2024

作者： Rajarajeswari, P. Madhubala, R. Gomathi, V. Ramco Institute of Technology Department of Computer Science and Engineering Rajapalayam India National Engineering College Department of Computer Science and Engineering Kovilpatti India

ISBN: (纸本)9798350389609

Virtual Reality (VR) technology in health-care has emerged as a valuable tool for advancing diagnostic techniques and enhancing patient care. This research explores the application of VR in attention profile assessments, with a particular focus on diagnosing Attention Deficit/Hyperactivity Disorder (ADHD). ADHD poses challenges in maintaining attention, and traditional diagnostic approaches often lack real-world relevance. This study examines the efficacy of Attention Contours - a VR-based Continuous Performance Task (CPT) - in comparison to its conventional 2D (two-dimensional) counterpart for assessing attention profiles. Attention Contours VR incorporates varied virtual settings and language-neutral stimuli to improve ecological validity. User trials reveal greater engagement and fewer distractions in the VR environment, highlighting its potential for more accurate attention assessment. Participants showed a strong preference for the VR format, praising its immersive experience and enhanced usability. The findings emphasize the impact of VR in advancing attention assessment tools and suggest its promising role in ADHD assessment. © 2024 IEEE.

关键词： Virtual environments

来源：评论

学校读者我要写书评

暂无评论

Brain Tumor Retrieval in MRI Images with Integration of Optimal Features

引用

Journal of Harbin institute of technology(New Series) 2024年第6期31卷 71-83页

作者： N V Shamna B Aziz Musthafa Department of Computer Science and Engineering PA College of EngineeringMangaluru 574153KarnatakaIndia Department of Computer Science and Engineering Bearys Institute of TechnologyMangaluru 5754199KarnatakaIndia

This paper presents an approach to improve medical image retrieval, particularly for brain tumors, by addressing the gap between low-level visual and high-level perceived contents in MRI, X-ray, and CT scans. Traditional methods based on color, shape, or texture are less effective. The proposed solution uses machine learning to handle high-dimensional image features, reducing computational complexity and mitigating issues caused by artifacts or noise. It employs a genetic algorithm for feature reduction and a hybrid residual UNet(HResUNet) model for Region-of-Interest(ROI) segmentation and classification, with enhanced image preprocessing. The study examines various loss functions, finding that a hybrid loss function yields superior results, and the GA-HResUNet model outperforms the HResUNet. Comparative analysis with state-of-the-art models shows a 4% improvement in retrieval accuracy.

关键词： medical images brain MRI machine learning feature extraction and reduction content-based image retrieval(CBIR)

来源：评论

学校读者我要写书评

暂无评论

Evaluating Deep Learning Algorithms for MRI-Based Brain Tumor Classification

Evaluating Deep Learning Algorithms for MRI-Based Brain Tumo...

引用

2024 International Conference on Emerging Innovations and Advanced Computing, INNOCOMP 2024

作者： Singh, Gurpreet Chhabra, Amit Mittal, Ajay Department of Computer Science and Engineering Chandigarh India University Institute of Engineering and Technology Panjab University Department of Computer Science and Engineering Chandigarh India

ISBN: (纸本)9798350376470

The classification of brain tumors has significant importance in the realm of clinical diagnosis and the implementation of appropriate treatment strategies. The process of diagnosing a brain tumor is time-consuming and strongly dependent on the expertise and experience of the radiologist. The volume of data that necessitates processing has seen a significant surge in parallel with the growth in patient population, rendering previous protocols financially burdensome and inefficient. Numerous researchers have conducted investigations on a range of algorithms aimed at the detection and classification of brain tumors, with a focus on achieving high levels of accuracy and computational efficiency. In recent times, there has been a surge in the utilization of Deep Learning (DL) methodologies to construct automated systems that possess the ability to effectively diagnose or segment brain tumors within a reduced timeframe. Deep learning (DL) facilitates the use of a pre-existing Convolutional Neural Network (CNN) model designed for the analysis of medical pictures, with a special focus on the classification of brain tumors. The rapid advancement of Deep Learning techniques in the context of computer vision tasks has been made possible by the availability of extensive training datasets and advancements in model architecture, leading to enhanced approximations within a supervised learning framework. The growth has mostly been motivated by the need to implement these ideas. The use of deep learning methods has shown potential in enhancing the accuracy of brain tumor identification and classification via the utilization of magnetic resonance imaging (MRI). According to the findings, CNN is a superior technique that greatly boosts the accuracy by 93.4% and it is maximum as compared to other methods. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Bi-SMOTE: a novel framework for handling imbalanced datasets using machine learning techniques

引用

International Journal of Information technology (Singapore) 2025年第1期17卷 431-445页

作者： Tigga, Onima Pal, Jaya Mustafi, Debjani Department of Computer Science & amp Engineering Birla Institute of Technology Jharkhand Mesra Ranchi India

Working with Imbalance data in real-world problems is not so easy due to the different cardinality of classes. Several machine learning Techniques have been used to overcome this kind of problem for 100% original data for classification. In this research work, the proposed Bi-SMOTE approach on the imbalanced Red wine and Glass datasets using machine learning techniques, Random Forest (RF), K Nearest Neighbor (KNN), Logistic Regression (LR), and Decision Tree (DT) have been used. Moreover, preprocessing steps are done by Binarization and standardization. The accuracy and F1-score are calculated of without SMOTE / with SMOTE / proposed Bi-SMOTE approaches with classifiers RF, KNN, LR, and DT for 95% & 100% of original datasets. Here the Bi-SMOTE approach gives better results up to 12%, and up to 22% for accuracy and up to 29% and up to 28% for F1-score for the Red wine and the Glass dataset respectively. Moreover, mean squared error (MSE) is continuously decreased in these machine learning techniques with original data with SMOTE and Bi-SMOTE. Thus, the proposed approach Bi-SMOTE yields the highest values in all performance measures, including accuracy, and F1-score for 95% data than 100% data. © Bharati Vidyapeeth's institute of computer Applications and Management 2024.

关键词： Accuracy Binarization Class Imbalance Machine learning (ML) SMOTE (Synthetic minority over-sampling technique)

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：