检索结果-内蒙古大学图书馆

Artificial Intelligence and Smart systems (ICAIS), International conference on

作者： V. Srilakshmi Anupama Anumolu M. Aruna Safali Vallabhaneni Siva Parvathi Department of CSE DVR & Dr.HS MIC College of Technology Vijayawada AP India Department of Mathamatics DVR & Dr.HS MIC College of Technology Vijayawada AP India Department of CSE Dhanekula Institute of Engineering and Technology Vijayawada AP India Department of Information Technology Prasad V. Potluri Siddhartha Institute of Technology Vijayawada AP India

Alzheimer’s disease (AD) is the most common disease that can cause a brain disorder in a human aged above 65. Detecting and diagnosing AD becomes a more complicated and complex task by using various manual processes. DL and ML algorithms are most widely used to analyze the complex features from the medical data used to detect AD from various samples. Several types of sample formats are used to detect AD. This paper mainly focused on detecting the AD from the retinal fundus images. Analyzing the early symptoms of AD can prevent the patient’s life from permanent eye loss. ML algorithms are having various drawbacks that use complex computations and more computation time for the processing of data. The AD prediction is done by using the fundus color images collected from the Kaggle dataset. ML follows various steps to complete the task such as training, pre-processing and algorithm implementation. In the existing approaches, a limited number of parameters are used. Another disadvantage of the traditional algorithms shows the low accuracy and unmatched results. This paper introduced the hybrid-layered framework is developed to detect the AD from the fundus images dataset. Several performance metrics such as precision, recall, F1-score, and accuracy are used to show the results.

关键词： Training Measurement Manuals Learning (artificial intelligence) Prediction algorithms Retina optical imaging

来源：评论

学校读者我要写书评

暂无评论

深度学习策略下光纤中超短脉冲非线性传输过程表征及控制研究进展

引用

中国激光 2023年第11期50卷 162-170页

作者：隋皓朱宏娜贾焕玉欧洺余李祺罗斌邹喜华西南交通大学物理科学与技术学院成都四川610031 西南交通大学信息科学与技术学院成都四川610031

常规数值求解方法在表征光纤中超短脉冲的非线性传输过程时存在计算量大、效率低等局限。随着人工智能的快速发展,深度学习技术展现出了强大的计算能力、广泛的适用范围、良好的硬件移植性,在光纤中超短脉冲非线性传输过程表征和控制研... 详细信息

常规数值求解方法在表征光纤中超短脉冲的非线性传输过程时存在计算量大、效率低等局限。随着人工智能的快速发展,深度学习技术展现出了强大的计算能力、广泛的适用范围、良好的硬件移植性,在光纤中超短脉冲非线性传输过程表征和控制研究中具有巨大潜力。本文概述了深度学习技术及其在预测光纤中超短脉冲传输、超短脉冲重构及参数估计方面的研究进展,同时展望了深度学习与光纤中超短脉冲非线性传输这一新兴交叉技术的发展方向和挑战。

关键词：光纤光学超短激光脉冲传输非线性薛定谔方程光纤非线性效应深度学习

来源：评论

学校读者我要写书评

暂无评论

In-Memory Massive MIMO Linear Detector Circuit with Extremely High Energy Efficiency and Strong Memristive Conductance Deviation Robustness

In-Memory Massive MIMO Linear Detector Circuit with Extremel...

引用

IEEE conference on Global Communications (GLOBECOM)

作者： Jia-Hui Bi Shaoshi Yang Ping Zhang Sheng Chen School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing China School of Electronics and Computer Science University of Southampton Southampton U.K

ISBN: (数字)9798350351255

ISBN: (纸本)9798350351262

The memristive crossbar array (MCA) has been successfully applied to accelerate matrix computations of signal detection in massive multiple-input multiple-output (MIMO) systems. However, the unique property of massive MIMO channel matrix makes the detection performance of existing MCA-based detectors sensitive to conductance deviations of memristive devices, and the conductance deviations are difficult to be avoided. In this paper, we propose an MCA-based detector circuit, which is robust to conductance deviations, to compute massive MIMO zero forcing and minimum mean-square error algorithms. The proposed detector circuit comprises an MCA-based matrix computing module, utilized for processing the small-scale fading coefficient matrix, and amplifier circuits based on operational amplifiers (OAs), utilized for processing the large-scale fading coefficient matrix. We investigate the impacts of the open-loop gain of OAs, conductance mapping scheme, and conductance deviation level on detection performance and demonstrate the performance superiority of the proposed detector circuit over the conventional MCA-based detector circuit. The energy efficiency of the proposed detector circuit surpasses that of a traditional digital processor by several tens to several hundreds of times.

关键词： Performance evaluation Fading channels Operational amplifiers Power demand Detectors Massive MIMO Mean square error methods Energy efficiency Robustness Signal detection

来源：评论

学校读者我要写书评

暂无评论

Fast Real-World Implementation of a Direction of Arrival Method for Constrained Embedded IoT devices 22

Fast Real-World Implementation of a Direction of Arrival Met...

引用

Proceedings of the 12th International conference on the Internet of Things

作者： Tiago Troccoli Juho Pirskanen Aleksandr Ometov Jari Nurmi Ville Kaseva Faculty of Information Technology and Communication Sciences Tampere University Finland and Wirepas Finland Wirepas Finland Faculty of Information Technology and Communication Sciences Tampere University Finland

ISBN: (纸本)9781450396653

Direction of arrival (DOA) methods are found in many applications, and in the case of the Internet of Things (IoT), it is used for indoor localization. However, the implementation of DOA in IoT devices poses a real challenge, since they are computationally expensive complex numerical methods that could easily lead to resource starvation, unacceptable execution time, and rapid depletion of batteries of small constrained embedded systems typically found in IoT networks. This paper contributes to alleviating that problem, it presents a fast low-power optimized version of a DOA method called Unitary TLS ESPRIT. The optimization exploits the radio communication system design to avoid two time-consuming executions of eigendecomposition, and instead, it applies two simple Power Method algorithms. The result is a lightweight version of ESPRIT that can attain sub-millisecond execution time. To prove the solution’s viability, we carried out experiments on energy consumption, memory footprint, accuracy, and execution time for three floating-point formats in a commercial constrained embedded IoT device series without any operating system and software layers. Experiments show the solution satisfies the hardware requirements and the floating-point precision fully operated by the Floating-Point Unit is found to be the best option.

关键词： direction of arrival embedded systems ESPRIT signal processing

来源：评论

学校读者我要写书评

暂无评论

Nonlinear Distortion Mitigation via Coherent All-optical Reservoir Computing for Long-Haul IM-DD transmission systems

Nonlinear Distortion Mitigation via Coherent All-Optical Res...

引用

Asia Communications and Photonics conference and Exhibition (ACP)

作者： Guanju Peng Yaping Liu Zheng Li Kunpeng Zhu Zhiqun Yang Janping Li Shigui Zhang Zhanhua Huang Lin Zhang Key Laboratory of Opto-electronic Information Technology of Ministry of Education and Tianjin Key Laboratory of Integrated Opto-electronics Technologies and Devices School of Precision Instruments and Opto-electronics Engineering Tianjin University Tianjin China HMN Technologies Co. Ltd Tianjin China Peng Cheng Laboratory Shenzhen China

Photonic reservoir computing (RC), as a promising paradigm of optical signal processing, has attracted a great deal of attention in fiber nonlinearity equalization recently. For real long-haul transmission scenarios, we propose an all-optical RC operated in a coherent scheme to mitigate the nonlinear distortion of high-speed signals. Compared with the direct decision architecture, the proposed RC can dramatically improve the $\mathrm{Q}^{2}$ factor by 2.6 $\text{dB}$ for a single-wavelength system and 1.9 $\text{dB}$ for a 7-channel wavelength-division multiplexing (WDM) system over 3960-km fiber transmission. Furthermore, the robustness of the coherent all-optical RC is also verified by conducting studies on different training sequences. The coherent all-optical RC shows great potential in real-time nonlinearity mitigation for long-haul fiber transmissions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Tidal Crowds: A Federated Crowd Flow Prediction Algorithm 7

Tidal Crowds: A Federated Crowd Flow Prediction Algorithm

引用

2024 7th International conference on Geoinformatics and Data Analysis, ICGDA 2024

作者： Geçer, Melike Garbinato, Benoit Department of Information Systems University of Lausanne Vaud Lausanne Switzerland

ISBN: (纸本)9798400716249

Understanding crowd dynamics is critical for public safety, transportation optimization, and disease control. Federated learning, a privacy-preserving machine-learning approach, suggests to train models on the devices where data is generated. This maintains server independence for latency-sensitive applications and ensures data privacy as data never leaves the device. Additionally, federated learning facilitates load balancing among participants. Unlike traditional machine learning solutions, federated learning aggregates model parameters of all participating entity. Thus, aggregation in federated learning is challenging. In this paper, we propose Tidal Crowds, a federated crowd flow prediction algorithm, leveraging the benefits of federated learning. Tidal Crowds comes in three parts, data acquisition and processing, machine learning algorithm, and aggregation algorithm. The machine learning algorithm is based on DeepSTN+ that is capable of handling long-range spatial dependencies. Aggregation algorithms include FedAvg and FedProxy. Tidal Crowds aims to examine and enhance prediction accuracy while providing server independence and preserving privacy using federated learning. In our experiments, Tidal Crowds outperforms the original DeepSTN+ in various settings. © 2024 Copyright held by the owner/author(s).

关键词： crowd flow prediction federated learning machine learning mobility

来源：评论

学校读者我要写书评

暂无评论

Lightweight Edge AI Model for Real-Time Vessel Detection in Maritime Surveillance Using Multi-Modal Data

Lightweight Edge AI Model for Real-Time Vessel Detection in ...

引用

Mobile Computing and Sustainable Informatics (ICMCSI), International conference on

作者： S.Devika Priyadharshini K. Vadivazhagan Department of Computer and Information Science Annamalai University Tamilnadu India

ISBN: (数字)9798331522667

ISBN: (纸本)9798331522674

For the purpose of monitoring, it is necessary to design vessel detecting system that are not only accurate but also lightweight, efficient and capable of functioning on edge devices. Even if they are achieved, traditional deep learning methods require a significant amount of computational power, which makes it difficult to implement them in situations with limited resources, such as unmanned aerial vehicles (UAV) or remote marine monitoring stations as examples. A revolutionary lightweight Edge artificial intelligence method was created for real time vessel detection is presented in this work. Utilizing method architecture such as Mobile Net and Tiny YOLO, the advanced method compression technique such as pruning and quantization is applied to reduce the amount of computational complexity and memory footprint without affecting the accuracy of detection. Synthetic Aperture Rader (SAR) imaging and multi modal data sources such as optical and infrared images are employed into the proposed system to improve its robustness in a variety of maritime circumstances. It is further optimized for low latency operation which guarantees high speed inference which is essential for time sensitive applications such as identification of illicit fishing, the prevention or piracy and the search and rescue operations. It has been demonstrated through experimental evaluations on high resolution SAR dataset that the method is capable of delivering improved detection performance while simultaneously reducing the number of computational resources required. This result in a significant enhancement in both the speed of processing and energy efficiency. This finding provides the possibility of deploying this lightweight method on edge devices, which would revolutionize maritime surveillance by providing solution that are real time, dependable and scalable.

关键词： Adaptation models Accuracy Computational modeling Surveillance Image edge detection Soft sensors Autonomous aerial vehicles Real-time systems Robustness Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

Cloud-Edge Collaborative Human Activity Recognition System Design and Implementation

Cloud-Edge Collaborative Human Activity Recognition System D...

引用

International conference on Wireless Communications and Signal processing (WCSP)

作者： Xichen Zhang Chen Zhu Zhiguo Shi Polytechnic Institute Zhejiang University Hangzhou China College of Information Science and Electronic Engineering Zhejiang University Hangzhou China Key Laboratory of Collaborative Sensing and Autonomous Unmanned Systems of Zhejiang Province

This paper proposes an edge computing architecture for Human Activity Recognition (HAR) based on Raspberry Pi and HaaS100, aimed at enhancing real-time performance and system robustness. The proposed approach employs Edge-End inference within the communication range of edge nodes, while relying on Cloud Edge inference otherwise, thereby establishing a comprehensive Cloud Edge End system that leverages the capabilities of a mobile communication 5G private network Wi-Fi. To achieve accurate activity recognition, three algorithms, namely Convolutional Neural Network (CNN), CNN-Long Short-Term Memory (CNN-LSTM), and Residual Network-based CNN (ResNet-based CNN), are implemented and their performances are rigorously compared using Actitracker datasets. The experimental results demonstrate that the ResNet-based CNN exhibits superior performance, achieving an impressive F1-score of 0.977. Additionally, to address the challenge of model size and resource-constrained edge devices, model compression is employed through channel-wise data collection, effectively reducing the parameter size from 220,000 to 160,000, with a negligible 0.5% decrease in accuracy. Finally, the compressed model is deployed on edge devices, and retraining is conducted using self-collected datasets to ensure practicality and adaptability in real-world scenarios. The proposed HAR edge computing architecture showcases significant advancements in real-time performance, system robustness, and resource optimization, making it a promising solution for activity recognition in edge computing environments.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Visual SLAM for Dynamic Environments Based on Static Key-Points Detection

Visual SLAM for Dynamic Environments Based on Static Key-Poi...

引用

Virtual Reality (ICVR), IEEE International conference on

作者： Liangtao Zhong Xiaoqiang Zhang Lingyan Ran Yamin Han Hongyu Chu School of Information Engineering Southwest University of Science and Technology Mianyang China School of Computer Science Northwestern Polytechnical University Xi’an China College of Information Engineering Northwest A&F University Xianyang China

Recently, the technique of virtual reality has drawn great attention from both academics and industry fields. In many VR applications, the time-varying orientations and the positions of the VR headset are needed, where the technique of simultaneous localization and mapping (SLAM) can be applied. While great progress has been made in vision-based SLAM algorithms and systems in the recent two decades, the problem of visual SLAM in dynamic environments with moving objects is still unaddressed. In this paper, a static key-point detection approach is proposed based on optical flow feature clustering and potential moving region detection and is utilized for RGB-D SLAM in a dynamic environment. Firstly, considering the fact that there could be multiple moving objects in the scene, an optical flow feature is designed to cluster the multiple motion patterns in the scene. Secondly, an ultra-lightweight object detection network is utilized to detect potential moving regions in the scene, without using GPU resources. Finally, by utilizing the key-point clustering and potential moving regions detection, the static key-points are detected and inputted into the keypoint-based visual SLAM methods. The proposed approach is easy to insert as a pre-processing module in various key-point based visual SLAM systems. Extensive experimental results with qualitative and quantitative evaluations have demonstrated the performance with satisfactory accuracy of the proposed approach in various challenging dynamic scene scenarios.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Circular Synthetic Aperture Radar for Near-field Terahertz Imaging

Circular Synthetic Aperture Radar for Near-field Terahertz I...

引用

International conference on Multimedia, Signal processing and Communication Technologies, (IMPACT)

作者： Mavis Gezimati Ghanshyam Singh Department of Electrical and Electronic Engineering Science Centre for Smart Information and Communication Systems University of Johannesburg Johannesburg South Africa

ISBN: (纸本)9781665476485

The potential of synthetic aperture radar (SAR) systems to yield high resolution images has led to its exploration in new applications. The conventional microwave radar imaging suffer limited range resolution in the sub-millimeter range and optical imaging methods are constrained by the diffraction limit. In the Terahertz (THz) spectrum, SAR systems are capable of operating in all weather, all time and penetrating through clouds, smoke, dust etc. to achieve better resolution images beyond the diffraction limit. In this paper, the different aspects of Multimedia computing are investigated in THz circular SAR as a new mode of radar imaging for near-field applications for example indoor environment. We have explored a new toolbox that enables the rapid development of near-field THz imaging SAR systems and generation of large near-field THz imaging scenarios that could be used for data driven applications. A circular SAR imaging system is presented for 2D (two-dimensional) imaging of targets in THz-SAR and the radar images are reconstructed using the Back Propagation and Polar Formatting algorithms. The SAR performance evaluation metrics are reported and scope for future work is discussed.

关键词： Terahertz wave imaging Image resolution Imaging Signal processing algorithms Multimedia computing Radar imaging Reconstruction algorithms

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：