检索结果-内蒙古大学图书馆

Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW), IEEE conference on

作者： Jerald Thomas Mohammed Safayet Arefin Doug Bowman University of Wisconsin - Milwaukee Colorado State University Virginia Tech

ISBN: (数字)9798331514846

ISBN: (纸本)9798331525637

Video-pass-through (VPT) augmented reality (AR) enables greater field of view and fidelity compared to current optical-see-through AR technologies. However, the computational complexity of VPT AR has traditionally resulted in high latency, low resolution, and/or visual artifacts. Only recently have devices come on the market that successfully mitigate these drawbacks, such as the Varjo XR-3 and XR-4, the Apple Vision Pro, and the Meta Quest 3. We propose that low-latency VPT AR introduces several interesting 3D User Interface (3DUI) capabilities. In this position paper, we discuss one of these 3DUIs, namely stereo-correct digital magnification of the physical environment. Specifically, this paper presents a novel system for stereo-correct digital magnification of the physical environment using the Varjo XR-3 (but that can be generalized to other VPT AR systems). The digital magnification algorithm we propose enables stereoscopic magnification of physical objects in real time by utilizing the front-facing camera feeds of the VPT AR system, implemented entirely through software. In addition, we outline the challenges and limitations associated with the proposed system, discuss our current progress, and suggest potential future directions for the development of our proposed VPT AR digital magnification system.

关键词： Performance evaluation Visualization Three-dimensional displays Stereo image processing Software algorithms User interfaces Software Real-time systems Usability Augmented reality

来源：评论

学校读者我要写书评

暂无评论

Development of a Novel Plantar Pressure Insole and Inertial Sensor System for Daily Activity Classification and Fall Detection 16th

Development of a Novel Plantar Pressure Insole and Inertial ...

引用

16th International conference on Intelligent Robotics and Applications, ICIRA 2023

作者： Fan, Bingfei Yi, Fugang Yang, Shuo Du, Mingyu Cai, Shibo The Key Laboratory of Special Purpose Equipment and Advanced Processing Technology Ministry of Education and Zhejiang Province Hangzhou China College of Mechanical Engineering Zhejiang University of Technology Hangzhou310014 China

ISBN: (纸本)9789819964857

Quantifying human daily activities can provide relevant monitoring information about physical activities and fall risk, and wearable sensors are promising devices for activity monitoring in daily life scenarios. This paper designed a novel plantar pressure insole and inertial sensor system and presented classification algorithms for activity classification and fall detection. We designed each plantar pressure insole with eight thin uniaxial load cells placed in the key area of a foot. Twenty healthy young adults performed selected activities in the laboratory while wearing the plantar pressure shoes and six inertial measurement units (IMUs) on their feet, shanks, and thighs of both sides. We adopted the convolutional neural network (CNN), ensemble learning, and support vector machine (SVM) methods for activity classification, and the input data were inertial data, pressure data, and both data. We adopted CNN, RNN (recurrent neural network), LSTM (long short-term memory), and CNN-LSTM method for fall detection, and compared results before and after the Dempster-Shafer evidence theory. Results show that for activity classification, CNN with both inertial and plantar pressure data got the best accuracy of 97.1%. For fall detection, the accuracy of RNN, CNN, LSTM, and CNN-LSTM were 93.77%, 95.85%, 96.16%, and 97.76%, respectively. LSTM got comparable accuracy as CNN-LSTM but with much less latency. The presented wearable system and algorithms show good feasibility in activity classification and fall detection, which could serve as a foundation for physical activity monitoring and fall alert systems for elderly people. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Design and Performance Evaluation of Multispectral Sensing algorithms on CPU, GPU, and FPGA

Design and Performance Evaluation of Multispectral Sensing A...

引用

IEEE Aerospace conference (AeroConf)

作者： Menon, Vivek V. Siddiqui, Saquib A. Rao, Sanil Schmidt, Andrew French, Matthew Chirayath, Ved Li, Alan Univ Southern Calif Informat Sci Inst Los Angeles CA 90007 USA NASA Ames Lab Adv Sensing Mountain View CA USA

ISBN: (纸本)9781728174365

Multispectral, Imaging, Detection, and Active Reflectance (MiDAR) is a method capable of imaging through ocean waves without distortion in 3D at sub-cm resolutions to sense living and non-living structures in light-limited and analog planetary science environments. MiDAR utilizes high intensity narrowband structured optical radiation to illuminate a target and measure the object's spectral reflectance at fine and spatial temporal scales with a signal-to-noise ratio 10 - 10(3) higher than passive airborne and spaceborne remote sensing systems. The MiDAR data is then reconstructed using the computational imaging capability associated with airborne platform such as unmanned aerial vehicles (UAVs) and spaceborne platforms such as a CubeSat. UAVs deploy embedded Graphics processing Unit (GPU) such as Nvidia's Tegra, while CubeSat consists of Field Programmable Gate Array (FPGA) devices such as Xilinx's MPSoC. The challenges involved in mapping MiDAR to each platform is unique and this paper describes the implementation and optimization steps to achieve maximum performance from each accelerator using CUDA and High Level Synthesis (HLS). The FPGA implementation using HLS is 14.7x faster than the CUDA-based GPU implementation and also provides a higher throughput of 13,552 fps using Xilinx MPSoC as compared to 921 fps achieved using Nvidia Tegra. Multispectral, Imaging, Detection, and Active Reflectance (MiDAR) is a method capable of imaging through ocean waves without distortion in 3D at sub-cm resolutions to sense living and non-living structures in light-limited and analog planetary science environments. MiDAR utilizes high intensity narrowband structured optical radiation to illuminate a target and measure the object's spectral reflectance at fine and spatial temporal scales with a signal-to-noise ratio 10 - 10(3) higher than passive airborne and spaceborne remote sensing systems. The MiDAR data is then reconstructed using the computational imaging capability associated

关键词： Reflectivity Performance evaluation Instruments Graphics processing units optical distortion optical variables measurement Throughput

来源：评论

学校读者我要写书评

暂无评论

CSI-Based Calibration Free Localization with Rotating Antenna for Coal Mine 16th

CSI-Based Calibration Free Localization with Rotating Antenn...

引用

16th International conference on Wireless algorithms, systems, and Applications (WASA)

作者： Zhang, Tieyang Zhang, Kuiyuan Liu, Dongjingdian Chen, Pengpeng China Univ Min & Technol Xuzhou Peoples R China

ISBN: (纸本)9783030859282;9783030859275

Accurate mine localization has always been an urgent demand. With the widespread emergence of cheap wireless devices in our life, channel state information (CSI) based indoor localization has attracted people's attention. In the existing CSI-based localization methods, fingerprint-based localization needs to build a huge database with poor robustness, and the methods based on angle of arrival (AOA) need specific antenna arrays and complex phase calibration. Because mobile devices such as smartphones only have one or two antennas, and the antennas inside the devices cannot form a receiving array, the conventional AOA estimation system cannot be deployed on such mobile devices. Therefore, we propose an AOA estimation method using rotating AP antennas. This method only needs one antenna at the receiving end to work and does not need to calibrate the phase. Firstly, the CSI phase of the rotating antenna is collected and the phase difference is calculated. Then, empirical mode decomposition (EMD) algorithm based on mutual information is used to remove the noise, and subcarrier selection is used to reduce the multipath effect in the data processing part. Finally, the AOA is estimated according to the relative position of the antenna. We conduct extensive experiments in mine and laboratory scenarios, and the median localization errors in mine and laboratory are 0.4m and 0.6 m, the median angle errors are 4.2 degrees and 5.2 degrees respectively.

关键词： Indoor localization Channel state information Smartphone

来源：评论

学校读者我要写书评

暂无评论

Tracking Quantum Coherence in Polariton Condensates with Time-Resolved Tomography

引用

Physical Review Letters 2023年第11期130卷 113601-113601页

作者： Carolin Lüders Matthias Pukrop Franziska Barkhausen Elena Rozas Christian Schneider Sven Höfling Jan Sperling Stefan Schumacher Marc Aßmann Experimentelle Physik 2 Technische Universität Dortmund D-44221 Dortmund Germany Department of Physics and Center for Optoelectronics and Photonics Paderborn (CeOPP) Universität Paderborn 33098 Paderborn Germany Institute of Physics University of Oldenburg D-26129 Oldenburg Germany Technische Physik Physikalisches Institut and Würzburg-Dresden Cluster of Excellence ct.qmat Universität Würzburg 97074 Würzburg Germany Theoretical Quantum Science Institute for Photonic Quantum Systems (PhoQS) Paderborn University Warburger Straße 100 33098 Paderborn Germany Wyant College of Optical Sciences University of Arizona Tucson Arizona 85721 USA

Long-term quantum coherence constitutes one of the main challenges when engineering quantum devices. However, easily accessible means to quantify complex decoherence mechanisms are not readily available, nor are sufficiently stable systems. We harness novel phase-space methods—expressed through non-Gaussian convolutions of highly singular Glauber-Sudarshan quasiprobabilities—to dynamically monitor quantum coherence in polariton condensates with significantly enhanced coherence times. Via intensity- and time-resolved reconstructions of such phase-space functions from homodyne detection data, we probe the systems’ resourcefulness for quantum information processing up to the nanosecond regime. Our experimental findings are confirmed through numerical simulations, for which we develop an approach that renders established algorithms compatible with our methodology. In contrast to commonly applied phase-space functions, our distributions can be directly sampled from measured data, including uncertainties, and yield a simple operational measure of quantum coherence via the distribution’s variance in phase. Therefore, we present a broadly applicable framework and a platform to explore time-dependent quantum phenomena and resources.

关键词： Exciton polariton Quantum information with hybrid systems Quantum interference effects Quantum optics Resource theories Semiconductor quantum optics Polariton condensate Semiconductor microcavities Homodyne & heterodyne detection Phase space dynamics Phase space methods

来源：评论

学校读者我要写书评

暂无评论

Single-scan Run Length Algorithm for Real-time Fiducial Marker Detection Based on FPGA devices 22

Single-scan Run Length Algorithm for Real-time Fiducial Mark...

引用

Proceedings of the 2022 6th International conference on Video and Image processing

作者： Farhana Binte Sufi Fernando Rincón Calle Jesús Barba Romero Juan Carlos López López Department of Electrical and Electronic Engineering University of Rajshahi Bangladesh Department of Technologies and Information Systems University of Castilla-La Mancha Spain

ISBN: (纸本)9781450397568

Image processing pipelines involved in detecting markers from real-time video employ contour detection algorithms. Field Programmable Gate Arrays (FPGA) can provide greater computing power due to parallel processing and pipelining of the different stages in these algorithms. However, FPGA implementation of different contour detection methods remains challenging due to resource limitation and timing constraints. In this work a novel approach for a complete pipeline has been presented based on FPGA. The method used for the detection of fiducial markers from real-time video streams is based on a single-scan approach. Implementation of such pipeline into reconfigurable technology is specially challenging compared to software-oriented implementation as pointers, dynamic structures, efficient use of resources, etc. need to be considered. A driver is the need to fit the design into low-cost reconfigurable devices. To help face the related challenges of a hardware implementation High Level Synthesis (HLS) technology has been adopted, which implies reengineering of the reference algorithms. The results show a successful detection of markers over a dataset acquired in a controlled indoor environment with different resolutions at different distances. The proposal has been prototyped and tested on a Zedboard with Xilinx Zynq®-7000 SoC.

关键词： Single-scan run length algorithm Contour detection algorithm FPGA Stream processing Edge computer vision Fiducial markers HLS

来源：评论

学校读者我要写书评

暂无评论

Fetal QRS Detection from Single-Channel Abdominal ECG by Adaptive Improved Clustering

Fetal QRS Detection from Single-Channel Abdominal ECG by Ada...

引用

International Workshop on Wearable and Implantable Body Sensor Networks (BSN)

作者： Thinh Nguyen-Quang Tai Le Duc Nguyen Minh Hung Cao Huy-Dung Han School of Electrical and Electronic Engineering Hanoi University of Science and Technology Hanoi Vietnam Department of Biomedical Engineering University of California Irvine Irvine CA USA Department of Electrical Engineering and Computer Science University of California Irvine Irvine CA USA

ISBN: (数字)9798331530143

ISBN: (纸本)9798331530150

Assessment of fetal development and wellness throughout pregnancy is essential in the early detection of pregnancy anomalies. Recent advances in wearable technology have led to the development of home pregnancy monitoring systems, which hold potentials for the continuous monitoring of expectant mothers and their unborn children through fe-tal/maternal electrocardiogram (f/mECG) technology. Given the complexity of extracting useful information from f/mECG data, these devices are often equipped with multiple channels and sophisticated signal processing algorithms. Thus, the implementation of f/mECG monitoring in daily life remains unfeasible. In this study, we propose an algorithm for extracting the fetal electrocardiogram (fECG) using a single channel. The algorithm detects the peaks of the maternal ECG (mECG) through a clustering method, followed by peak correction. The mECG signal is then reconstructed with Principal Component Analysis (PCA) before applying the Template Subtraction (TS) method to obtain fECG. Rigorous testing with various online f/mECG databases demonstrates the potential to accurately extract fECG using only a single channel. The evaluation results based on the online dataset achieved an accuracy of 99.60% and an F1 score of 99.30%. This advancement also opens opportunities to optimize home-based f/mECG monitoring, making it more compact and reliable.

关键词： Pregnancy Measurement Accuracy Fetal heart rate Signal processing algorithms Clustering algorithms Filtering algorithms Electrocardiography Monitoring Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

A Transformer-Block-Wise Collaborative Training Mechanism with Hybrid Parallelism Over Heterogeneous Networks

A Transformer-Block-Wise Collaborative Training Mechanism wi...

引用

IEEE conference on Wireless Communications and Networking

作者： Jiewei Chen Jingrong Wang Shaoyong Guo Jiakai Hao Xuesong Qiu Zehui Xiong State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications State Grid Beijing Information & Communication Company Pillar of Information Systems Technology and Design Singapore University of Technology and Design

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

With the rise of AI-Generated Content (AIGC) services in wireless networks, efficient and high-quality distributed training of Large Language Models (LLMs) has become essential for enabling the large-scale application of next generation AI technologies. However, the extensive parameters of LLMs impose significant demands on memory, computing power and communication resources in heterogeneous networks. To efficiently utilize the dispersed network resources, this paper presents a First-Pipeline- Then-Federated Learning (FPTFL) approach with a hybrid parallel scheduling strategy to facilitate the training of Transformer-based LLMs. We propose a block-wise splitting mechanism to partition the Transformer's encoder into distinct segments, which are deployed cross individual devices. The encoder parameters and intermediate smashed data are uploaded to the edge server, where the whole model is updated through federated aggregation. Particularly, we develop a fine-grained computation-efficient method based on pipeline parallelism, enabling the segments to cooperatively train the entire encoder. An optimization problem is formulated to determine the LLM segments and the number of micro-batches under network resource constraints, with the goal of minimizing the total latency of LLM training services. Simulation results demonstrate that our approach enables Transformer-based model training on resource-constrained devices, preserves model performance, and reduces waiting time.

关键词： Training Performance evaluation Large language models Wireless networks Pipelines Parallel processing Transformers Heterogeneous networks Classification algorithms Optimization

来源：评论

学校读者我要写书评

暂无评论

An investigation on all-optical multicast technology in all-optical networks 2021

An investigation on all-optical multicast technology in all-...

引用

3rd International conference on Robotics systems and Automation Engineering, RSAE 2021

作者： Ma, Yuqi College of Information Technology Beijing University of Technology China

ISBN: (纸本)9781450388467

optical fiber communication has the characteristics of large transmission capacity, long transmission distance and good transmission efficiency, which could satisfy the ever-increasing bandwidth data communication requirements. All-optical network can directly process signal in the optical domain with the conventional optical - electrical - optical conversion and improve the signal processing speed. All-optical multicast technology is an important part of all-optical networks, which can improve the flexibility of the network and make better use of network resources. In this paper, four kinds of nonlinear effects are investigated to realize all-optical multicast, including cross-gain modulation (XGM), cross-phase modulation (XPM), four-wave mixing (FWM) and cascaded sum frequency generation (SFG)/difference frequency generation (DFG). The principle and multicast performance of these nonlinear effects are discussed. Moreover, the characteristics of several kinds of nonlinear devices are also compared, including SOA, HNLF, PPLN. © 2021 ACM.

关键词： Four wave mixing

来源：评论

学校读者我要写书评

暂无评论

Real-Time Anomaly Detection in 5G Networks Through Edge Computing

Real-Time Anomaly Detection in 5G Networks Through Edge Comp...

引用

IEEE International conference on Intelligent Techniques in Control, Optimization and Signal processing (INCOS)

作者： Riaz Shaik Dara Raju Prakash Chandra Behera Ravindra Changala S. Suma Christal Mary A. Balakumar Department of CSE KL University Vaddeswaram Andhra Pradesh India Dept. of Computer Science & Engineering Vignana Bharathi Institute and Technology Aushapur Ghatkesar India Department of Science St. Claret College Bangalore India Department of Information Technology Guru Nanak Institutions Technical Campus Hyderabad India Department of Information Technology Panimalar Engineering College Poonamalle Chennai Department of ECE K. Ramakrishnan College Of Engineering samayapuram Trichy India

ISBN: (数字)9798350361186

ISBN: (纸本)9798350361193

Strong anomaly detection techniques are becoming more and more necessary as 5G networks develop in order to maintain network performance, security, and dependability. This study leverages the capabilities of Mobile Edge Computing (MEC) to present a novel method for anomaly detection in 5G networks. By processing data closer to the network edge, the integration of MEC offers an efficient and decentralized architecture that lowers latency and improves real-time detection capabilities. The distributed module takes advantage of its close proximity to network devices by using sophisticated algorithms for anomaly detection, which are implemented at the mobile edge. The system can quickly detect abnormalities from typical network activity by utilizing capabilities including Flow Collection, Anomaly Symptom Detection, and Network Anomaly Detection. The distributed module provides anomalous information to the centralized decision-making module for thorough examination. It takes into account variables like resource use and network traffic and integrates this data with metrics gathered from monitoring modules. Because of its adaptive characteristics, the system may expand anomaly detection components, enhance detection functions, and modify virtualized resources in response to shifting network circumstances. The evaluation findings reveal that the suggested anomaly detection method performs well in 5G networks, with decreased false positives, increased responsiveness, and better flexibility to changing network conditions. Using MEC not only makes anomaly detection more effective, but it also fits in with the 5G design, which makes it a viable option for protecting the upcoming generation of communication networks. It obtains 95% accuracy in classification. The suggested approach has proven to be resilient in handling security issues by producing outcomes that are either comparable to or better than those attained by other techniques that have been previously presented in the s

关键词： Multi-access edge computing 5G mobile communication Image edge detection Signal processing algorithms Telecommunication traffic Signal processing Real-time systems

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：