检索结果-内蒙古大学图书馆

A Use of DL Method with the 5G Integration for image Analysis which is in Underwater with Aerial Vehicle kind of Networks 1

A Use of DL Method with the 5G Integration for Image Analysi...

引用

1st International Conference on Innovative Sustainable Technologies for Energy, Mechatronics and Smart Systems, ISTEMS 2024

作者： Agrawal, Anant Prakash Krishna Reddy, B.N. Malviya, Ashwini Noida Institute of Engineering and Technology Department of Mechanical Engineering Uttar Pradesh Greater Noida India Department of Management Karnataka Bangalore India Atlas SkillTech University Department of Ugdx Maharashtra Mumbai India

ISBN: (纸本)9798350384246

deep neural networks, also referred to as DNNs, have experienced great progress in processing a wide range of data types, including pictures, time series, speech, audio, and video, thanks to their amazing ability to infer representations from data. Within the discipline of remote sensing, surveys and studies of the literature devoted to DNN applications are being conducted in an attempt to compile the massive amount of data produced in subfields. Applications using unmanned aircraft are a common theme in aerial sensing research. Nevertheless, there hasn't been a thorough literature study that addresses the issues of 'deep learning' and 'unmanned aerial vehicles remotely sensing' as of yet. In order to close this gap, this work proposes a novel approach for underwater drone vehicle networks that makes use of deep learning (DL) techniques for obtaining and categorising features in picture analysis. Pictures taken by unmanned aerial vehicles (UAVs) outfitted with a This module undergo a series of pre-processing procedures, such as noise reduction, smoothing, and normalisation. The modified pictures are then subjected to feature extraction using convolutional neural networks based on multilayer extreme learning. Radial basis functions networks based on recursive elimination are then used to classify the deep characteristics that were retrieved. The performance of the suggested technique is assessed by experimental analysis using several UAV image datasets, with a focus on precision, recall, precision, and accuracy The F measure, root-median-square error (RMSE), and average precision of the mean (MAP). © 2024 IEEE.

关键词： Radial basis function networks

来源：评论

学校读者我要写书评

暂无评论

Efficientnet-Based deep learning Approach for Video Forgery Detection and Authentication 10

Efficientnet-Based Deep Learning Approach for Video Forgery ...

引用

10th International Conference on Communication and Signal processing, ICCSP 2024

作者： Gowri Priyaa, V. Jaya Harrish, M. Udhayakumar, M. Jothieswaran, N. Dinesh, K. Coimbatore Institute of Technology Electronics and Communication Engineering Coimbatore India

ISBN: (纸本)9798350353068

The technological advancements in artificial intelligence have made it a lot easier to create forged videos that are difficult to distinguish from reality. Fake videos also called deep fakes are created with greater accuracy and precision. Detecting and removing fake data on the internet can prevent misinformation and rumors from spreading. To achieve this, detection methods must be robust, generalized, fast, and accurate enough to detect fake data. In this paper, deepfakes are created using Generative Adversarial Network (GAN) and used for dataset training. The deepfakes are found to be different from real ones by various parameters like facial expressions, irregularities in the image, etc. This project focuses on detecting the manipulated face of the person in a frame using the EfficientNet B4 algorithm. The EfficientNet B4 model is more accurate than EfficientNet B0 and less complex than EfficientNet B7. The modified EfficientNet B4 outperforms the existing EfficientNet B0 in terms of accuracy. The probabilities of deep fakes in each frame are calculated and on average the video is detected as real or fake. This model demonstrates a very successful detection rate of more than 92%. Finally, modified EfficientNet B4 is compared with other models' performance. © 2024 IEEE.

关键词： Discriminators

来源：评论

学校读者我要写书评

暂无评论

Data anomaly detection with automatic feature selection and deep learning

引用

STRUCTURES 2023年 57卷

作者： Jiang, Huachen Ge, Ensheng Wan, Chunfeng Li, Shu Quek, Ser Tong Yang, Kang Ding, Youliang Xue, Songtao Southeast Univ Key Lab Concrete & Prestressed Concrete Struct Minist Educ Nanjing 210096 Peoples R China Natl Univ Singapore Dept Civil Engn Singapore 119596 Singapore Tsinghua Univ Hefei Inst Publ Safety Res Hefei 230601 Peoples R China Tongji Univ Res Inst Struct Engn & Disaster Reduct Coll Civil Engn Shanghai 200092 Peoples R China Tohoku Inst Technol Dept Architecture Sendai Miyagi 9828577 Japan State Key Lab Safety Durabil & Hlth Operat Long Sp Nanjing 210012 Peoples R China Key Lab Rd & Bridge Detect & Maintenance Technol R Hangzhou 310023 Peoples R China

As increasing sensor-based structural health monitoring (SHM) systems are implemented on civil infrastructures, sensor data reliability plays a crucial role in the assessment of operational performance of bridges. Sensors are heavily exposed to harsh environmental conditions during their operations and inevitably lead to possible unstable performance or failure. Thus, to accurately identify faulty sensors is a prerequisite to processing and analyzing the collected data for assessment purpose. Recently, researchers adopted the convolutional neural network (CNN) approach to identify faulty sensors, focusing on image features. Such approach may overlook some important detailed signal features and the time series approach may still be needed. However, algorithms based on time series tend to be time consuming because of the lengthy and high dimensional dataset. This may be effectively resolved using an automatic feature selection technique, namely Tsfresh, as proposed in this paper to select highly relevant signal features based on statistical tests of significance. A deep learning technique based on fully convolutional network (FCN) can then be efficiently employed for anomaly classification. The algorithm is validated using a dataset collected from a real cable-stayed bridge and results show that the proposed method significantly reduces the training time for the neural network, albeit with high classification accuracy.

关键词： Structural health monitoring Anomaly detection Convolutional neural network Feature selection Tsfresh deep learning

来源：评论

学校读者我要写书评

暂无评论

Efficient LEO Satellite Data Management Using No-Reference image Quality Assessment 30

Efficient LEO Satellite Data Management Using No-Reference I...

引用

30th IEEE International Conference on Telecommunications, ICT 2024

作者： Sboui, Nourhane Ghazzai, Hakim Najeh, Sameh Sboui, Lokman University of Carthage Tunis Tunisia Thuwal Saudi Arabia Montreal Canada

ISBN: (纸本)9798350356694

Low Earth Orbit (LEO) satellite communications for Internet-of-Things (IoT) services provide a cost-effective means to improve existing networks. However, challenges related to data, such as storage capacity, bandwidth consumption and image transfer speed, necessitate the development of real-time data management solutions for LEO satellite communications. In this paper, we introduce a comprehensive no-reference deep image quality assessment (NR-DIQA) framework designed to effectively manage LEO satellite images for Earth observation (EO) by identifying and filtering distorted and anomalous images before transmission. The proposed approach employs a labeling methodology for remote sensing images (RSI) with a deep learning (DL) model to compute a score that assesses image quality (IQ) and determines the set of high-quality images to be transmitted to the ground station. We show that the proposed approach overcomes a baseline model and show that extensive data savings could be achieved for EO applications. © 2024 IEEE.

关键词： Information management

来源：评论

学校读者我要写书评

暂无评论

Enhanced Electroacoustic Tomography with Supervised learning for real-time Electroporation Monitoring

引用

PRECISION RADIATION ONCOLOGY 2024年第3期8卷 110-118页

作者： Jiang, Zhuoran Xu, Yifei Sun, Leshan Srinivasan, Shreyas Wu, Q. Jackie Xiang, Liangzhong Lei, Ren Stanford Univ Stanford CA USA Univ Calif Irvine Irvine CA 92697 USA Duke Univ Med Ctr Durham VA USA Univ Maryland Sch Med Baltimore MD 21250 USA

Background: Nanosecond pulsed electric fields (nsPEF)-based electroporation is a new therapy modality potentially synergized with radiation therapy to improve treatment outcomes. To verify its treatment accuracy intraoperatively, electroacoustic tomography (EAT) has been developed to monitor in-vivo electric energy deposition by detecting ultrasound signals generated by nsPEFs in real-time. However, utility of EAT is limited by image distortions due to the limited-angle view of ultrasound transducers. Methods: This study proposed a supervised learning-based workflow to address the ill-conditioning in EAT reconstruction. Electroacoustic signals were detected by a linear array and initially reconstructed into EAT images, which were then fed into a deep learning model for distortion correction. In this study, 56 distinct electroacoustic data sets from nsPEFs of different intensities and geometries were collected experimentally, avoiding simulation-to-real-world variations. Forty-six data were used for model training and 10 for testing. The model was trained using supervised learning, enabled by a custom rotating platform to acquire paired full-view and single-view signals for the same electric field. Results: The proposed method considerably improved the image quality of linear array-based EAT, generating pressure maps with accurate and clear structures. Quantitatively, the enhanced single-view images achieved a low-intensity error (RMSE: 0.018), high signal-to-noise ratio (PSNR: 35.15), and high structural similarity (SSIM: 0.942) compared to the reference full-view images. Conclusions: This study represented a pioneering stride in achieving high-quality EAT using a single linear array in an experimental environment, which improves EAT's utility in real-time monitoring for nsPEF-based electroporation therapy.

关键词： electroacoustic tomography electroporation interventional therapy limited-angle reconstruction supervised learning

来源：评论

学校读者我要写书评

暂无评论

Research on YOLOv8-Based Household Item Detection for Indoor Scenes 1

Research on YOLOv8-Based Household Item Detection for Indoor...

引用

1st IEEE International Conference on Data Intelligence and Innovative Application, DIIA 2024

作者： Wu, Jingwen Luo, Xiaonan Guilin University of Electronic Technology Guangxi Key Laboratory of Image and Graphic Intelligent Processing Guilin China

ISBN: (纸本)9798331504434

This research is driven by the growing need for efficient, scalable object detection in smart home applications with real-time performance and resource constraints. In this study, we utilize the YOLOv8 deep learning model for household item detection, focusing on its performance in real-time scenarios. We compare different versions of YOLOv8 - specifically, YOLOv8n, YOLOv8s, and YOLOv8m - using a preprocessed custom dataset. The results indicate that YOLOv8n stands out due to its lower parameter count and faster inference speed, making it ideal for real-time applications, even though it has slightly lower precision and recall. In contrast, while YOLOv8s and YOLOv8m show improvements in certain metrics, their higher computational demands limit their effectiveness in resource-constrained settings. Overall, YOLOv8n effectively balances speed and accuracy, offering a viable solution for household item detection. © 2024 IEEE.

关键词： Smart homes

来源：评论

学校读者我要写书评

暂无评论

Research on Automatic Detection and Early Warning of Epilepsy in Electroencephalogram Signals 9th

Research on Automatic Detection and Early Warning of Epileps...

引用

9th International Conference on Internet of Things, ICIOT 2024, Held as Part of the Services Conference Federation, SCF 2024

作者： Zheng, Shu-Xiong Li, Si-Tong Zhang, Hui-Lin Bao, Juan School of Public Health Hubei University of Medicine Shiyan442000 China

ISBN: (纸本)9783031770029

With the enhancement of computing capabilities at edge computing nodes, edge computing based on cloud computing has seen widespread development and application. In the field of epilepsy prevention and treatment, applications related to edge computing have become very popular. Edge computing is a distributed computing model that places data processing, storage, and application functions at the edge location close to the data source, enabling real-time data processing, reducing data transmission latency, and improving system response speed. The temporal complexity of epilepsy electroencephalogram (EEG) signals has been described and characterized at different time periods. Addressing the multi-channel, high-dimensional, and heterogeneous nature of EEG signals and considering their uncertainty and dynamic characteristics, by analyzing and mining foundational data and combining deep learning theories and methods, epilepsy patients’ EEG signals are analyzed and differentiated across multiple time periods, delving into the characteristics of epilepsy EEG signals at different times. In traditional EEG data collection and management systems, cloud storage faces challenges when dealing with large amounts of terminal data, which can reduce real-time data processing performance. To address these issues, this paper proposes an EEG wireless data collection and analysis system based on edge computing. By pushing data processing and storage capabilities to the network edge, data can be processed and analyzed in real-time at the source, thereby reducing data transmission delays and network congestion. Building upon this, the focus is on automatic identification, prediction, and decision-making related to epilepsy, exploring the dynamic patterns of epilepsy automatic identification processes, breaking through traditional epilepsy automatic identification, prediction, and decision-making models, proposing a new method for epilepsy automatic identification, prediction, and decision-ma

关键词： deep learning

来源：评论

学校读者我要写书评

暂无评论

GDRMA: Graph Neural Networks for Document Retrievals With Mean Aggregation

引用

IEEE ACCESS 2024年 12卷 185706-185727页

作者： Maya, Shigeru Toshiba Corp Res & Dev Ctr Syst AI Lab Kawasaki Kanagawa 2128582 Japan

With the proliferation of cloud services and high-capacity hard drives, the volume of stored document data is rapidly increasing. Consequently, large-scale document retrieval tasks have been attracting significant attention. Recently, embedding-based methods including language models and graph neural networks (GNNs) have been developed to effectively handle synonyms in documents. However, a major limitation of these approaches is scalability. When taking N-grams into account, it is important to remember that many query keywords are unsupported by language models and that existing GNN-based methods can cause GPU memory shortages. To address this issue, we propose Graph neural networks for Document Retrievals with Mean Aggregation (GDRMA). First, we carefully select a subset of words as important words and derive document embeddings using our novel GNNs on the important words-documents graph to save GPU memory usage. Then, we quickly learn an embedding of the target query keyword using "mean aggregation" and generate a ranking of related documents on CPUs. The main advantage is that our provided GNN connects the two steps mentioned above smoothly, and the generated ranking incorporates synonyms based on a co-occurrence relationship. We conducted exhaustive experiments on real datasets and confirmed that GDRMA is superior to comparable methods.

关键词： deep learning Batch production systems real-time systems Graph neural networks Thesauri Long short term memory Graphics processing units Computational modeling Bipartite graph Anomaly detection Document retrieval embedding graph graph neural network (GNN) network recommendation

来源：评论

学校读者我要写书评

暂无评论

Research on deep learning Based Exercise Instruction Methods 2

Research on Deep Learning Based Exercise Instruction Methods

引用

2nd International Conference on Signal processing and Intelligent Computing, SPIC 2024

作者： Wang, Jiaye Liu, Yunting Ding, Haifeng Shenyang Ligong University School of Automation and Electrical Engineering Shenyang China

ISBN: (纸本)9798350368888

Due to the lack of fitness exercise foundation, leisure time, the limitations of the sports ground, and other factors. A fitness exercise guidance system is designed to help people exercise effectively. In this paper, the improved OpenPose model is used to collect the skeletal information of the human body, and the collected information is normalized. The pose vector cosine distance is introduced to represent the single-frame similarity. The motion sequences are compared with the standard sequences, and the dynamic regularization algorithm is used to calculate the overall similarity and provide real-time feedback on motion correction. Finally, the overall network is applied to the baduanjin fitness exercise, which has the significance of promotion. © 2024 IEEE.

关键词： Dynamic time Warping MobileNetV2 OpenPose

来源：评论

学校读者我要写书评

暂无评论

UAV-Enhanced Service Caching for IoT Systems in Extreme Environments

引用

IEEE INTERNET OF THINGS JOURNAL 2024年第16期11卷 26741-26750页

作者： Yan, Hanzhi Li, Hanwen Xu, Xiaolong Bilal, Muhammad Nanjing Univ Informat Sci & Technol Sch Comp Sci Nanjing 210044 Peoples R China Nanjing Univ Informat Sci & Technol Sch Software Nanjing 210044 Peoples R China Nanjing Univ Informat Sci & Technol Jiangsu Collaborat Innovat Ctr Atmospher Environm Nanjing 210044 Peoples R China Hankuk Univ Foreign Studies Dept Comp Engn Yongin 17035 Gyeonggi South Korea

The proliferation of Internet of Things (IoT) applications and real-time services brings severe performance pressure on IoT systems with cloud computing, so edge computing is increasingly being adopted in IoT systems to assist cloud computing in providing services. Systems with cloud-edge computing deploy parts of services on edge servers located closer to IoT devices, thus enabling real-time data processing and analysis and improving the Quality of Experience (QoE) of users. However, inevitable extreme events (e.g., meteorological disasters) and the aging of the physical infrastructure cause varying degrees of performance impairment to edge servers, which adversely affects the service provisioning capability of IoT systems. Therefore, there is a serious challenge to cope with the lack of service provisioning capability owing to the impaired edge server performance in extreme environments. In this article, an unmanned aerial vehicle (UAV) -assisted service provisioning framework for the IoT systems in cloud-edge computing is introduced, and a UAV-enhanced service caching scheme based on a potential game (G-USC) is proposed for this framework. Besides, to provide a prerequisite for service caching, a UAV position update scheme based on a deep $Q$ -network is designed. The experimental analysis proves that G-USC effectively solves the problem of insufficient service provisioning capability of edge servers in extreme environments.

关键词： Internet of Things Servers Cloud computing Edge computing Autonomous aerial vehicles real-time systems Game theory Cloud-edge computing game theory Internet of Things (IoT) reinforcement learning service caching unmanned aerial vehicles (UAVs)

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：