检索结果-内蒙古大学图书馆

Deep Trajectory Classification Model for Congestion Detection in Human Crowds

computers, Materials & Continua 2021年第7期68卷 705-725页

作者： Emad Felemban Sultan Daud Khan Atif Naseer Faizan Ur Rehman Saleh Basalamah Department of Computer Engineering College of Computing and Information SystemsUmm Al-Qura UniversityMakkahSaudi Arabia Department of Computer Science National University of TechnologyIslamabadPakistan Science and Technology Unit Umm Al-Qura UniversityMakkahSaudi Arabia Institute of Consulting Research and Studies Umm Al-Qura UniversityMakkahSaudi Arabia

In high-density gatherings,crowd disasters frequently occur despite all the safety *** detection of congestion in human crowds using automated analysis of video footage can prevent crowd *** work on the prevention of crowd disasters has been based on manual analysis of video *** methods also measure crowd congestion by estimating crowd ***,crowd density alone cannot provide reliable information about *** paper proposes a deep learning framework for automated crowd congestion detection that leverages pedestrian *** proposed framework divided the input video into several temporal *** then extracted dense trajectories from each temporal segment and converted these into a spatio-temporal image without losing information.A classification model based on convolutional neural networks was then trained using spatio-temporal ***,we generated a score map by encoding each point trajectory with its respective class *** this,we obtained the congested regions by employing the non-maximum suppression method on the score ***,we demonstrated the proposed framework’s effectiveness by performing a series of experiments on challenging video sequences.

关键词： Crowd congestion trajectory classification crowd analysis

来源：评论

学校读者我要写书评

暂无评论

MotionAGFormer: Enhancing 3D Human Pose Estimation with a Transformer-GCNFormer Network

MotionAGFormer: Enhancing 3D Human Pose Estimation with a Tr...

引用

IEEE Workshop on Applications of computer Vision (WACV)

作者： Soroush Mehraban Vida Adeli Babak Taati KITE Research Institute University of Toronto Institute of Biomedical Engineering University of Toronto Department of Computer Science University of Toronto

Recent transformer-based approaches have demonstrated excellent performance in 3D human pose estimation. However, they have a holistic view and by encoding global relationships between all the joints, they do not capture the local dependencies precisely. In this paper, we present a novel Attention-GCNFormer (AGFormer) block that divides the number of channels by using two parallel transformer and GCNFormer streams. Our proposed GCNFormer module exploits the local relationship between adjacent joints, outputting a new representation that is complementary to the transformer output. By fusing these two representation in an adaptive way, AGFormer exhibits the ability to better learn the underlying 3D structure. By stacking multiple AGFormer blocks, we propose MotionAGFormer in four different variants, which can be chosen based on the speed-accuracy trade-off. We evaluate our model on two popular benchmark datasets: Human3.6M and MPI-INF-3DHP. MotionAGFormer-B achieves state-of-the-art results, with P1 errors of 38.4 mm and 16.2 mm, respectively. Remarkably, it uses a quarter of the parameters and is three times more computationally efficient than the previous leading model on Human3.6M dataset. Code and models are available at https://***/TaatiTeam/MotionAGFormer.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Privacy-preserving Human Activity Recognition via Video-based Range-Doppler Synthesis

Privacy-preserving Human Activity Recognition via Video-base...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Zhiyuan Cui Luoyu Mei Siyuan Pei Borui Li Xiaolei Zhou School of Computer Science and Engineering Southeast University Department of Computer Science City University of Hong Kong The Sixty-Third Research Institute National University of Defense Technology

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

As an important branch of IoT applications, Human activity recognition (HAR) is widely used in daily life, particularly through vision-based methods. However, vision-based HAR has serious privacy issues. How to better and low-cost protect the privacy of users who have already installed the relevant devices is a problem that needs to be solved. To address this challenge, we can solve it by transforming video to privacy-preserving mmWave data. Existing studies have primarily focused on synthesizing micro-Doppler data from video, but there is a lack of methods for synthesizing range-Doppler data. Thus, we present a comprehensive method for synthesizing range-Doppler data from videos and subsequently utilize this synthetic data for HAR. Experimentally, we deploy our range-Doppler synthesis method and classification model on a custom dataset. Experimental results indicate that the model trained with synthetic data achieves accuracy on the custom dataset by 95.7%, which is comparable to the accuracy of vision-based HAR works, and demonstrate that the scheme proposed in this paper achieves privacy-preserving HAR.

关键词： Privacy Accuracy Federated learning Data models Sensors Human activity recognition Internet of Things

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Survey on RIS-Enhanced Physical Layer Security in UAV-Assisted Networks

引用

IEEE Internet of Things Journal 2025年

作者： Ahmed, Manzoor Soofi, Aized Amin Raza, Salman Li, Yongxiao Khan, Feroz Khan, Wali Ullah Asif, Muhammad Han, Zhu School of Computer and Information Science China Hubei Engineering University Institute for AI Industrial Technology Research Xiaogan City432000 China National University of Modern Languages Department of Computer Science Faisalabad38000 Pakistan National Textile University Department of Computer Science Faisalabad38000 Pakistan Beijing University of Posts and Telecommunications School of Electronic Engineering Beijing China Gachon University School of Computing Seongnam-si13120 Korea Republic of School of Computer Science Department of Electrical and Computer Engineering United States University of Houston Department of Electrical and Computer Engineering HoustonTX77004 United States Kyung Hee University Department of Computer Science and Engineering Seoul446-701 Korea Republic of

This survey provides an in-depth examination of the role of reconfigurable intelligent surfaces (RIS) in enhancing physical layer security (PLS) within unmanned aerial vehicle (UAV)-assisted networks, which are essential for the secure and efficient operation of sixth-generation (6G) wireless communications. The study covers various types of RIS—passive, active, and hybrid—and their applications in both terrestrial and aerial environments to strengthen PLS. Key focus areas include advanced PLS techniques such as optimizing UAV trajectory, beamforming, and RIS phase-shift configurations, all aimed at improving secrecy rates (SRs) while mitigating the risks of eavesdropping and jamming. Moreover, the survey also addresses strategies for enhancing energy-efficient SRs and implementing anti-jamming mechanisms within UAV-assisted networks. Additionally, it explores the integration of RIS-UAV systems with emerging technologies such as non-orthogonal multiple access (NOMA), mobile edge computing (MEC), cognitive radio, and THz networks, demonstrating how security can be enhanced in such networks. Through detailed performance analysis, the paper highlights the transformative potential of RIS-equipped UAVs in overcoming the potential security challenges for future 6G networks. Finally, the survey presents lessons learned and identifies critical future research directions and open challenges, offering insights that will guide the development of robust and secure RIS-assisted UAV systems in next-generation wireless networks. © 2014 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Network-based analysis to uncover drug-induced adverse side-effects

Network-based analysis to uncover drug-induced adverse side-...

引用

2023 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2023

作者： Funari, Alessio Paci, Paola Conte, Federica Sapienza University of Rome Department of Translational and Precision Medicine Rome Italy National Research Council Institute for Systems Analysis and Computer Science 'Antonio Ruberti' Rome Italy Sapienza University of Rome Department of Computer Control and Management Engineering Rome Italy Karolinska Institutet Stockholm17177 Sweden

ISBN: (纸本)9798350337488

Despite the drug approval process consists of extremely rigorous clinical and preclinical studies, not all side effects are identified before its marketing, posing a significant risk to public health. Furthermore, considering the huge use of economic and human resources, in-silico predictive approaches for the identification of side effects are essential. In this study, we introduce a new method based on random walk with restart algorithm to delineate previously unidentified links between drugs and side effects, and we apply it on the drug-induced Asthma and long QT syndrome. We identified the genes potentially involved in the development of the analyzed side effect by comparing side-effect-related drugs with drugs not known to induce side effects. Analyzing the sets of genes most likely influenced by the perturbation of each individual drug, we observed that, on average, side-effect-related drugs perturb a higher percentage of genes involved in the development of side effects compared to side-effect-unrelated drugs. Based on this finding, we developed a classifier to explore all possible unknown associations between drugs and side effects. This method can be extended to the analysis of other side effects as well. © 2023 IEEE.

关键词： network medicine random walk side effect

来源：评论

学校读者我要写书评

暂无评论

Automated Detection of Manipulated Images Using Deep Learning and computer Vision Techniques

Automated Detection of Manipulated Images Using Deep Learnin...

引用

International Conference on Parallel Distributed and Grid computing (PDGC)

作者： Atika Gupta Priya Matta School of Computing Graphic Era Hill University Dehradun India Department of Computer Science and Engineering Tula’s Institute Dehradun India

ISBN: (数字)9798331521349

ISBN: (纸本)9798331521356

In the current era, where image manipulation techniques have developed at a faster pace, detection of altered images using the conventional methods has become increasingly challenging. This article presents a framework for automatically detecting the altered images using deep learning and computer vision techniques. The system focuses on providing methods to detect various image manipulations such as copy-move, inpainting and image-slicing. With the help of integrating convolutional neural networks (CNN) with anomaly detection methods and image preprocessing, the proposed framework can successfully find visual image anomalies that show potential image manipulation. The experiment conducted shows that the framework could effectively detect image alterations using the state-of-the-art computer vision and deep learning algorithms.

关键词： Deep learning computer vision Visualization Accuracy Computational modeling Splicing computer architecture Data augmentation Robustness Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Highly ordered inkjet-printed quantum-dot thin films enable efficient and stable QLEDs with EQE exceeding 23%

引用

escience 2024年第3期4卷 152-162页

作者： Changting Wei Bo Xu Meng Zhang Zhenhuang Su Jiawei Gu Wenrui Guo Xingyu Gao Wenming Su Zheng Cui Seokwoo Jeon Zhiyong Fan Haibo Zeng MIIT Key Laboratory of Advanced Display Materials and Devices Institute of Optoelectronics&NanomaterialsSchool of Materials Science and EngineeringNanjing University of Science and TechnologyNanjing 210094China Ningbo Nato New Material Technology Co. LtdNingbo 315600China Shanghai Synchrotron Radiation Facility(SSRF) Zhangjiang LaboratoryShanghai Advanced Research InstituteChinese Academy of SciencesShanghai 201204China School of Mechanical Engineering Jiangsu UniversityZhenjiang 212013China Printable Electronics Research Center Nano Devices and Materials DivisionSuzhou Institute of Nano-Tech and Nano-BionicsChinese Academy of SciencesSuzhou 215123China Department of Materials Science and Engineering Korea UniversitySeoul 02841South Korea Department of Electronic and Computer Engineering The Hong Kong University of Science and TechnologyHong Kong 999077China

Inkjet-printed quantum dot light-emitting diodes(QLEDs)are emerging as a promising technology for next-generation ***,the progress in fabricating QLEDs using inkjet printing technique has been slower compared to spin-coated devices,particularly in terms of efficiency and *** key to achieving high performance QLEDs lies in creating a highly ordered and uniform inkjet-printed quantum dot(QD)thin *** this study,we present a highly effective strategy to significantly improve the quality of inkjet-printed CdZnSe/CdZnS/ZnS QD thin films through a pressure-assisted thermal annealing(PTA)*** from this PTA process,a high quality QD thin film with ordered packing,low surface roughness,high photoluminescence and excellent electrical property is *** mechanism behind the PTA process and its profound impact on device performance have been thoroughly investigated and ***,a record high external quantum efficiency(EQE)of 23.08%with an impressive operational lifetime(T50)of up to 343,342h@100cdm−2,and a record EQE of 22.43%with T50 exceeding to 1,500,463h@100cdm−2 are achieved in inkjet-printed red and green CdZnSe-based QLEDs,*** work highlights the PTA process as an important approach to realize highly efficient and stable inkjet-printed QLEDs,thus advancing QLED technology to practical applications.

关键词： Inkjet printing Quantum dot light-emitting diode High efficiency Thermal annealing Highly ordered

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning for Job Scheduling and Resource Management in Cloud computing: An Algorithm-Level Review

arXiv

引用

arXiv 2025年

作者： Gu, Yan Liu, Zhaoze Dai, Shuhong Liu, Cong Wang, Ying Wang, Shen Theodoropoulos, Georgios Cheng, Long The School of Control and Computer Engineering North China Electric Power University Beijing China The School of Computer Science and Technology Shandong University of Technology China The Institute of Computing Technology Chinese Academy of Sciences Beijing China The School of Computer Science University College Dublin Ireland Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China

Cloud computing has revolutionized the provisioning of computing resources, offering scalable, flexible, and on-demand services to meet the diverse requirements of modern applications. At the heart of efficient cloud operations are job scheduling and resource management, which are critical for optimizing system performance and ensuring timely and cost-effective service delivery. However, the dynamic and heterogeneous nature of cloud environments presents significant challenges for these tasks, as workloads and resource availability can fluctuate unpredictably. Traditional approaches, including heuristic and meta-heuristic algorithms, often struggle to adapt to these real-time changes due to their reliance on static models or predefined rules. Deep Reinforcement Learning (DRL) has emerged as a promising solution to these challenges by enabling systems to learn and adapt policies based on continuous observations of the environment, facilitating intelligent and responsive decision-making. This survey provides a comprehensive review of DRL-based algorithms for job scheduling and resource management in cloud computing, analyzing their methodologies, performance metrics, and practical applications. We also highlight emerging trends and future research directions, offering valuable insights into leveraging DRL to advance both job scheduling and resource management in cloud computing. © 2025, CC BY.

关键词： Job shop scheduling

来源：评论

学校读者我要写书评

暂无评论

Scale-Aware Super-Resolution Network With Dual Affinity Learning for Lesion Segmentation From Medical Images

引用

IEEE Transactions on Neural Networks and Learning Systems 2024年第6期36卷 10530-10543页

作者： Luyang Luo Yanwen Li Zhizhong Chai Huangjing Lin Pheng-Ann Heng Hao Chen Department of Computer Science and Engineering The Hong Kong University of Science and Technology Sai Kung Hong Kong Department of Biomedical Informatics Harvard University Cambridge MA USA Imsight AI Research Lab Shenzhen China Department of Computer Science and Engineering The Chinese University of Hong Kong Shatin Hong Kong Department of Computer Science and Engineering Department of Chemical and Biological Engineering The Hong Kong University of Science and Technology Hong Kong China HKUST Shenzhen-Hong Kong Collaborative Innovation Research Institute Futian Shenzhen China

Convolutional neural networks (CNNs) have shown remarkable progress in medical image segmentation. However, the lesion segmentation remains a challenge to state-of-the-art CNN-based algorithms due to the variance in scales and shapes. On the one hand, tiny lesions are hard to delineate precisely from the medical images which are often of low resolutions. On the other hand, segmenting large-size lesions requires large receptive fields, which exacerbates the first challenge. In this article, we present a scale-aware super-resolution (SR) network to adaptively segment lesions of various sizes from low-resolution (LR) medical images. Our proposed network contains dual branches to simultaneously conduct lesion mask SR (LMSR) and lesion image SR (LISR). Meanwhile, we introduce scale-aware dilated convolution (SDC) blocks into the multitask decoders to adaptively adjust the receptive fields of the convolutional kernels according to the lesion sizes. To guide the segmentation branch to learn from richer high-resolution (HR) features, we propose a feature affinity (FA) module and a scale affinity (SA) module to enhance the multitask learning of the dual branches. On multiple challenging lesion segmentation datasets, our proposed network achieved consistent improvements compared with other state-of-the-art methods. Code will be available at: https://***/poiuohke/SASR_Net .

关键词： Lesions Image segmentation Feature extraction Medical diagnostic imaging Superresolution Diseases Decoding Computed tomography Technological innovation Skin

来源：评论

学校读者我要写书评

暂无评论

Next – Gen Driver Monitoring System

Next – Gen Driver Monitoring System

引用

Global Power, Energy and Communication Conference (GPECOM)

作者： Suganiya Murugan Harish Krishnadas S Abiishek G Pradeep Kumar S Rubini B Department of Computing Technologies SRM Institute of Science and Technology Chennai India Department of Computer Science and Engineering SRM Institute of Science and Technology Chennai India Department of Electrical and Electronics Engineering Vels Institute of Science Technology and Advanced Studies Chennai India

ISBN: (数字)9798350351088

ISBN: (纸本)9798350351095

the Next Gen Driver Monitoring System utilizes advanced techniques to enhance road safety by detecting driver drowsiness, nodding, and blink patterns. Integrating cutting-edge technology, the system analyzes facial landmarks and eye movements in real-time through sophisticated algorithms. Through the fusion of computer vision and advanced algorithms, it accurately identifies signs of driver fatigue and distraction, providing timely alerts to mitigate potential risks on the road. This research delves into the design, implementation, and evaluation of the system, demonstrating its effectiveness in reducing accidents caused by drowsy or inattentive driving. The Next Gen Driver Monitoring System represents a significant advancement in automotive safety technology, ushering in safer driving experiences.

关键词： computer vision Vehicle safety Fatigue Road safety Real-time systems Monitoring Vehicles

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：