检索结果-内蒙古大学图书馆

Multiple Pedestrian Tracking Under Occlusion: A Survey and Outlook

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2025年第2期35卷 1009-1027页

作者： Sun, Zhihong Wei, Guoheng Fu, Wei Ye, Mang Jiang, Kui Liang, Chao Zhu, Tingting He, Tao Mukherjee, Mithun Naval Univ Engn Dept Informat Secur Wuhan 430030 Hubei Peoples R China Wuhan Univ Sch Comp Sci Wuhan 430072 Peoples R China Suzhou Xingyu Technol Suzhou 215151 Peoples R China Harbin Inst Technol Sch Comp Sci & Technol Harbin 150001 Peoples R China Wuhan Univ Sch Comp Sci Wuhan 430072 Peoples R China Wuhan Univ Hubei Key Lab Multimedia & Network Commun Engn Wuhan 430072 Peoples R China Nanjing Univ Informat Sci & Technol Sch Artificial Intelligence Nanjing 210044 Peoples R China

As an intermediate task in computer vision, multiple pedestrian tracking (MPT) aiming at tracking the pedestrians from a given video, has attracted attention due to its potential academic and commercial value. However, pedestrians commonly suffer from occlusion due to diverse and complex scenarios, which increases the challenge of this task. This survey provides comprehensive review in terms of occlusion scenarios encountered during MPT, and investigates the model robustness of the existing methods in this scenarios. Firstly, this survey introduces the various and states of occlusion. Secondly, the related occlusion datasets are introduced. Subsequently, we categorize existing occlusion handling methods according to the tracking process and detail their pros and cons. In addition, occlusion handling precision (OHP) metric is proposed to evaluate the ability of a tracker in handling occlusion in this survey. Moreover, comprehensive analyzes and discussions in several public datasets are provided to verify the effectiveness of these methods. Finally, the existing issues and future directions for occlusion handling methods are discussed. In doing so, this work serves as a foundation for future research by providing researchers with information about the occlusion handling method of MPT.

关键词： Pedestrians Surveys Trajectory Target tracking Cameras Reviews Object tracking Circuits and systems Computer science computational modeling Multiple object tracking multiple pedestrian tracking occlusion handling trajectory prediction trajectory recovery

来源：评论

学校读者我要写书评

暂无评论

The Impact of Service Demand Variability on Data Center Performance

引用

IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 2025年第2期36卷 120-132页

作者： Olliaro, Diletta Anggraito, Adityo Marsan, Marco Ajmone Balsamo, Simonetta Marin, Andrea Univ Ca Foscari Venezia I-30123 Venice Italy IMDEA Networks Inst Leganes 28918 Spain Politecn Torino I-10129 Turin Italy

Modern data centers feature an extensive array of cores that handle quite a diverse range of jobs. Recent traces, shared by leading cloud data center enterprises like Google and Alibaba, reveal that the constant increase in data center services and computational power is accompanied by a growing variability in service demand requirements. The number of cores needed for a job can vary widely, ranging from one to several thousands, and the number of seconds a core is held by a job can span more than five orders of magnitude. In this context of extreme variability, the policies governing the allocation of cores to jobs play a crucial role in the performance of data centers. It is widely acknowledged that the First-In First-Out (FIFO) policy tends to underutilize available computing capacity due to the varying magnitudes of core requests. However, the impact of the extreme variability in service demands on job waiting and response times, that has been deeply investigated in traditional queuing models, is not as well understood in the case of data centers, as we will show. To address this issue, we investigate the dynamics of a data center cluster through analytical models in simple cases, and discrete event simulations based on real data. Our findings emphasize the significant impact of service demand variability, both in terms of requested cores and service times, and allow us to provide insight for enhancing data center performance. In particular, we show how data center performance can be improved thanks to the control of the interplay between service and waiting times through the assignment of cores to jobs.

关键词： Data centers Probability density function Data models Analytical models Time factors Queueing analysis computational modeling Standards Servers Resource management Data center performance analysis simulation multiserver job queuing model

来源：评论

学校读者我要写书评

暂无评论

Community Detection in Dynamic Networks: Exact Recovery Under Two Link Evolution Models

引用

IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 2025年第2期12卷 848-858页

作者： Moghaddam, Javad Zahedi Nosratinia, Aria Univ Texas Dallas Erick Johnson Sch Engn & Comp Sci Dept Elect & Comp Engn Richardson TX 75080 USA

This paper studies edge persistence and memory in time-varying (dynamic) networks, and their effect on community detection. In particular, we focus on two models representing network memory, using which we study the asymptotic behavior of community detection in dynamic networks of large size, manifested through their phase transition threshold. In the first part, we adopt a Markovian stochastic block model (SBM) in which the edge probabilities in each network snapshot depend not only on the respective node attributes but also on the previous network snapshot. Under this model, semi-definite relaxation achieves the optimal phase transition bound for exact recovery, as observed in other community detection problems. The adverse effect of edge persistence on perfect recovery is analyzed and highlighted. In the second part, we study networks where underlying communities change slowly compared with network measurements. We model this scenario via a time series of SBM wherein the node attributes are fixed within a window of a certain size and vary independently across windows. The phase transitions are calculated via semidefinite programming which, once again, is asymptotically optimal. Numerical simulations conducted on finite-size networks interpret the asymptotic results.

关键词： Image edge detection computational modeling Adaptation models Stochastic processes Sports Maximum likelihood detection Social networking (online) Semidefinite programming Numerical models Maximum likelihood estimation Community detection stochastic block model dynamic network maximum likelihood detection semidefinite programming edge persistence clustering

来源：评论

学校读者我要写书评

暂无评论

YOLORM: An Advanced Key Point Detection Method for Accurate and Efficient Rotameter Reading in Low Flow Environments

引用

IEEE ACCESS 2025年 13卷 1804-1816页

作者： Huang, Yong Xia, Xing Xiao, Shengwang Changsha Res Inst Min & Met CO Ltd Changsha 410006 Peoples R China

Automatic reading of rotameters in low flow and challenging environments poses substantial accuracy and efficiency challenges. To address these issues, this study introduces YOLORM, an advanced key point detection method for rotameters, built upon the YOLOv8n model. We de-veloped a comprehensive YOLO-RM dataset comprising 57,500 images to facilitate model training and evaluation. Several innovations were introduced to the YOLOv8n framework: (1) An ALKBlock-based C2f module incorporating DACov structure to enhance spatial perception;(2) An EffQA-FPN feature pyramid network inspired by QARepVGG to mitigate precision loss during quantization and reparameterization;(3) A DynamicHead attention mechanism with multi dimensional perception capabilities;and (4) EIoU and Adaptive Wing Loss functions to optimize bounding box and key point regression. Experimental results demonstrate that YOLORM achieved exceptional performance on the YOLO-RM dataset, with 99.67% precision, 98.83% recall, 99.48% mAP50, and 98.58% mAP50-95. Compared to the baseline YOLOv8n model, YOLORM achieved a 7.43% increase in detection accuracy and a 15.21% reduction in computational complexity. Moreover, YOLORM exhibited significant reductions in parameter count and computational cost while maintaining or enhancing detection performance relative to state-of-the-art algorithms. Notably, compared to Hourglass, HRNet, and HigherHRNet, YOLORM reduced parameters by 99.3%, 92.9%, and 96.8%, and computational cost by 98.8%, 90.2%, and 95.3%, respectively, while concurrently improving precision and recall. This study presents an efficient and reliable auto-matic reading solution for rotameters in industrial automation, demonstrating robust perfor-mance in complex environments. The proposed YOLORM model has the potential to significantly enhance safety and efficiency in industrial production processes.

关键词： Feature extraction Convolution Accuracy computational modeling YOLO Computer architecture Production Location awareness computational efficiency Adaptation models Industrial automation rotameter industrial measurement YOLORM model automatic reading method YOLO-RM dataset industrial automation

来源：评论

学校读者我要写书评

暂无评论

Multilevel and Energy-Efficient Partial Computation Offloading in Heterogeneous Edge Intelligence

引用

IEEE INTERNET OF THINGS JOURNAL 2025年第11期12卷 14993-15007页

作者： Xu, Baoyu Ruan, Yancheng Qiu, Chenghu He, Shuibing Shu, Feng Kang, Xiaoyang Zhang, Lihua Fudan Univ Acad Engn & Technol Shanghai 200433 Peoples R China Shanghai Univ Sch Comp Engn & Sci Shanghai 200444 Peoples R China Zhejiang Univ Coll Comp Sci & Technol Hangzhou 310027 Peoples R China Univ Sydney Sch Elect & Informat Engn Sydney NSW 2050 Australia Fudan Univ Acad Engn & Technol MOE Frontiers Ctr Brain Sci Lab Neural Interface & Brain Comp Interface Shanghai 200433 Peoples R China

Due to the diversity of edge devices (EDs) and applications, edge systems are heterogeneous and have been applied in artificial intelligence fields, such as smart factories and intelligent transportation, which is called heterogeneous edge intelligence. Many studies employ computation offloading to transfer processing data from resource-scarce EDs to resource-rich edge servers. These studies primarily focus on the overall resource consumption of homogeneous edge systems, neglecting the system heterogeneity and the details of resource consumption. In this article, we construct a system model from a parallel perspective for the heterogeneous edge system with different processors, memory, and applications, which perceives the cost of energy and delay from three levels: system, application, and component. A hybrid metaheuristic algorithm combined with a greedy rule, hybrid mutation, and whale optimization algorithm (GHMWOA) is proposed to realize partial computation offloading. A partial offloading architecture of heterogeneous edge intelligence is proposed to validate our model and algorithm with real-world hardware and software. Experiment results not only show GHMWOA outperforms multiple classical optimization algorithms in minimizing energy consumption, but also discover on which system component energy consumption depends, and how properties of application and system influence the cost of energy.

关键词： Delays Energy consumption Program processors Optimization Edge computing computational modeling Costs Artificial intelligence Servers Heuristic algorithms Heterogeneous edge intelligence multilevel energy analysis parallel processing partial computation offloading

来源：评论

学校读者我要写书评

暂无评论

DRANet: Deep Learning-Based Automatic White Balancing Approach to CVCC

引用

IEEE ACCESS 2025年 13卷 36714-36722页

作者： Choi, Ho-Hyoung Kyungpook Natl Univ Adv Dent Device Dev Inst Sch Dent Daegu 41940 South Korea

As a core component of the image processing pipeline, an image signal processor plays a critical role in automatic white balancing (AWB) for computer vision color constancy (CVCC). With the recent advent of deep convolutional neural network (DCNN), image signal processors have experienced enormous progress in CVCC. A myriad of deep learning-based CVCC models have emerged and outperformed their statistics-based, shallow counterparts. In this context, this article presents a novel deep learning-based AWB approach: the Dual Residual Aggregated Network (DRANet). Homogeneous Dual Residual Blocks (DRBs) are a key component of the proposed DRANet, intended to design the architecture with high cardinality. The homogeneous DRBs perform transformations simultaneously and their outputs are processed in a concatenating manner. This characterizes the proposed DRANet as a simple and wide but still deep structure which contributes to advancing illumination estimation accuracy to an innovative level and addressing the overfitting and long-term dependency issues that conventional approaches have long been struggling with. The proposed DRANet offers another advantage that comes with high cardinality, that is, a reduction in the number of parameters necessary to run the architecture. The experimental results provide compelling evidence that the proposed AWB approach has achieved an innovative progress in illumination estimation accuracy, as well as validating the invariance of both illumination and image device.

关键词： Image color analysis Computer architecture Accuracy Transforms Lighting Estimation Supervised learning computational modeling Convolutional neural networks Overfitting Deep learning-based CVCC DRANet illuminant estimation automatic white balancing dual residual block

来源：评论

学校读者我要写书评

暂无评论

Controlled Service Scheduling Scheme for User-Centric Software-Defined Network- Based Internet of Things

引用

IEEE ACCESS 2025年 13卷 19198-19218页

作者： Albekairi, Mohammed Jouf Univ Coll Engn Dept Elect Engn Sakakah 72388 Saudi Arabia

Software Defined Networks (SDNs) support different applications' data and control operations through operational plane differentiations. Such differentiations rely on the service providers' user density and processing capacity. This article introduces a Controlled Service Scheduling Scheme (CS3) to ensure responsive user service support. This scheme exploits the SDN's operation plane differentiation to confine immobile request stagnancies. The routed regression learning model decides the SDN plane selection. This learning is a modified version of linear learning where the scheduling rate is the plane differentiator. The process is un-iterated until the combination of device processing capacity and number of devices is less than the service population observed. In the scheduling process, the operation to data plane migrations is decided using the maximum routed threshold. The threshold is computed for the operation and data plane from which the rate of service response or capacity of service admittance is decided. The routed regression analyzes the change in the threshold factor to ensure flexible scheduling is achieved regardless of dense IoT requests. This scheme achieves a high scheduling rate for maximizing service distributions under controlled delay. The experimental findings show that compared to the current models, the suggested method improves the scheduling rate by 13.92%, increases the distribution of services by 8.31%, and decreases delays by 11.58%. Further evidence of the approach's efficacy in managing heavy IoT traffic is its low distribution failure rate of 1.7%. These findings demonstrate that the scheme can enhance performance in ever-changing Internet of Things settings by optimizing the allocation of resources.

关键词： Internet of Things Resource management Processor scheduling Job shop scheduling computational modeling Dynamic scheduling Delays Real-time systems Quality of service Software defined networking Control plane IoT regression learning SDN service scheduling

来源：评论

学校读者我要写书评

暂无评论

Fast prediction of curing residual stress in polymer-matrix composites by a deep learning method

引用

POLYMER COMPOSITES 2025年第0期 8424-8441页

作者： Zhang, Fan Guo, Haili Wang, Weiwei Du, Jinbo Liu, Yuchi Wang, Han Zhejiang Univ Sch Mech Engn State Key Lab Fluid Power & Mechatron Syst Hangzhou 310027 Peoples R China Zhejiang Univ Sch Mech Engn Key Lab Adv Mfg Technol Zhejiang Prov Hangzhou Peoples R China State Owned Wuhu Machinery Factory Wuhu Peoples R China Shanghai Acad Spaceflight Technol Shanghai Inst Satellite Engn Shanghai Peoples R China

Curing residual stress (CRS) is common in polymer-matrix composites due to the anisotropic properties of materials. Finite element method (FEM), the most extensively used approach for curing behavior prediction, is usually complicated and time-consuming. To achieve a fast prediction of process-induced stresses, a convolutional neural network (CNN) is established based on the FEM. Firstly, a fully coupled methodology is built and validated through distortions of experimentally manufactured laminates. Then, it is applied to generate models with different stacking layers, and the residual stress is computed and serves as the dataset of the deep learning model. Finally, the construction and hyperparameters are determined, and the good generalization performance proves the high accuracy of the current model. Besides, the CNN method (<1 s) greatly reduces the computational time compared with the FEM (>14 min). The supervised machine learning method shows great potential in promoting efficiency in stacking sequence designing and optimization of composites.

关键词： computational modeling cure behavior finite element analysis (FEA) polymer-matrix composites (PMCs)

来源：评论

学校读者我要写书评

暂无评论

Privacy-Preserving Outsourced Computation of Collaborative Operational Decisions Among Microgrids in an Active Distribution Network

引用

IEEE TRANSACTIONS ON POWER SYSTEMS 2025年第1期40卷 850-865页

作者： Han, Xutao Li, Zhiyi Xiao, Xuanyi Ju, Ping Shahidehpour, Mohammad Zhejiang Univ Coll Elect Engn Hangzhou 310027 Peoples R China IIT Robert W Galvin Ctr Elect Innovat Chicago IL 60616 USA

This article presents an operational decision-making scheme for facilitating the collaborative decisions between the utility distribution grid (UDG) and microgrids (MGs) in an active distribution network (ADN). The collaborative decision-making among UDG and MGs can help maximize the social welfare of ADN operations, but the decision-making process is faced with ADN computation timeliness and MG privacy leakage concerns. Hence, we propose a proactive deception approach for privacy-preserving outsourced computation of MGs that offers collaborative decisions of UDG and MGs in a centralized and performance-guaranteed manner. First, MG dispatch models are proactively transformed by constructing virtual MGs, which are outsourced to the distribution system operator (DSO), without changing their respective structures and sparsity. Then, the DSO integrates the encrypted MG dispatch models, including virtual objectives, variables, and parameters, with the UDG dispatch model to offer collaborative ADN decisions. Further, the designed asynchronous decryption mechanism will return the DSO's collaborative solutions about points of common coupling which are decrypted in parallel by MG master controllers for updating the distorted UDG power flow. Mathematically, the convexity, feasibility, and correctness of collaborative ADN decision-making are maintained by the proposed approach. Finally, through numerical experiments, the modified IEEE 33- and 123-bus systems are applied to demonstrate the rapidity, accuracy, robustness, and scalability of the proposed approach.

关键词： Collaboration computational modeling Biological system modeling Vectors Encryption Costs Decision making Active distribution network collaborative decision-making microgrids outsourced computation proactive deception

来源：评论

学校读者我要写书评

暂无评论

An Intelligent Framework for Cluster-Based Side-Channel Analysis on Public-Key Cryptosystems

引用

IEEE INTERNET OF THINGS JOURNAL 2025年第2期12卷 1962-1973页

作者： Wei, Congming He, Shulin Wang, An Sun, Shaofei Ding, Yaoling Zhang, Jingqi Zhu, Liehuang Beijing Inst Technol Sch Cyberspace Sci & Technol Beijing 100081 Peoples R China Beijing Inst Technol Sch Comp Sci & Technol Beijing 100081 Peoples R China Chinese Acad Sci Inst Informat Engn State Key Lab Informat Secur Beijing 100093 Peoples R China Beijing Inst Technol Sch Integrated Circuits & Elect Beijing 100081 Peoples R China

Classical cluster-based side-channel analysis (SCA) uses clustering algorithms to analyze power traces and often, principal component analysis to reduce the dimension of data, resulting in that clustering may not deal well with high-dimensional traces, such as cryptographic algorithm implementations with countermeasures. In this article, we propose an intelligent framework for cluster-based SCA, which includes three steps of clustering, classification and correction, for processing large high-dimensional data. By combining unsupervised clustering and supervised deep learning techniques, the framework succeeds in mining the data for additional in-depth information. In addition, unlike traditional cluster-based SCA, our approach focuses on deep learning and deliberately avoids over-reliance on cluster labels during classification. And metrics for correction are adopted to achieve a high level of reliability in key recovery. Experiments on the RSA smart card based on Montgomery ladder implementation and FPGA-based ECC with random delay demonstrate that our framework can significantly improve the success rate with strong robustness.

关键词： Clustering algorithms Deep learning Noise Convolutional neural networks Classification algorithms Cryptography computational modeling Cluster deep learning public-key cryptosystems side-channel analysis (SCA)

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：