检索结果-内蒙古大学图书馆

A center-driven image set partition algorithm for efficient structure from motion

INFORMATION SCIENCES 2019年 479卷 101-115页

作者： Sun, Kun Tao, Wenbing China Univ Geosci Sch Comp Sci Hubei Key Lab Intelligent Geoinformat Proc Wuhan 430074 Hubei Peoples R China Huazhong Univ Sci & Technol Sch Automat Natl Key Lab Sci & Technol Multispectral Informat Wuhan 430074 Hubei Peoples R China

This paper proposes a novel center-driven image set partitioning method dedicated for efficient Structure from Motion (SfM) on unevenly distributed images. First, multiple base clusters are found at places with high image density. Instead of building a small initial model from two images, we build multiple initial base models from these base clusters. This promises that the scene is reconstructed from dense places to sparse areas, which can reduce error accumulation when images have weak overlap. Second, the whole image set is divided into several region clusters to decide which images should be reconstructed from the same base model. In this step, the base models are treated as centers and the affinity between an image with each of them is measured by the reconstruction path length. To enable faster speed, images in each region cluster are further divided into several sub-region clusters so that they could be added to the same base model simultaneously. Based on the above partitioning results, the partial 3D models are reconstructed in parallel and then merged. Experiments show that the proposed method achieves remarkable speedup and better completeness than state-of-the-art methods, without significant accuracy deterioration. (C) 2018 Elsevier Inc. All rights reserved.

关键词： Center-driven image set partitioning 3D reconstruction Structure from Motion

来源：评论

学校读者我要写书评

暂无评论

parallel image and video self-recovery scheme with high recovery capability

Parallel image and video self-recovery scheme with high reco...

引用

conference on Real-Time image processing and Deep Learning

作者： Molina-Garcia, Javier Ponomaryov, Volodymyr, I Reyes-Reyes, Rogelio Cruz-Ramos, Clara Inst Politecn Nacl ESIME Culhuacan Ciudad De Mexico Mexico

ISBN: (纸本)9781510626584

In this paper, a parallel scheme for self-recovery of tampered images and videos is proposed. Designed technique is based on two methods for generating the digest image: halftoning and a block based scheme, additionally, the implementation of an authentication algorithm was carried out using a block-based method. In order to obtain robustness to the tampering coincidence problem, the proposed scheme embeds multiple versions of the recovery watermark. Finally, during the recovery process, an algorithm of inverse halftoning was applied to improve the quality of the reconstructed image. Proposed scheme was implemented in such a way that each process should be highly parallelizable using GPU, multicore processors. Experimental results have shown that novel framework generates a good quality of the watermarked images and the recovered images. The simulation results using parallel architectures have demonstrated the efficiency of the novel technique when it is implemented in a real-time environment.

关键词： parallel Computing GPU processing CPU multi-core CUDA image Self-Recovery Watermarking image Authentication

来源：评论

学校读者我要写书评

暂无评论

Optimizing Algorithms for low CPU Usage in Different Scenarios 20

Optimizing Algorithms for low CPU Usage in Different Scenari...

引用

20th International Carpathian Control conference (ICCC)

作者： Popa, Bogdan Popescu, Dan Roman, Monica Constantinescu, Radu Lucian Univ Craiova Dept Automat Control & Elect Craiova Romania

ISBN: (纸本)9781728107011

This paper proposes a series of results over the image processing algorithms and for the shortest path algorithm that influence the percentage of CPU utilization. Starting with the fractal generation to image analysis, compression for big images and also any algorithm, the parallel programing is today one of the most important factor for low consumption and good power balanced for the processing unit. A good balance between the processing capacity, the algorithm and the time of processing is hard to obtain due to the number of calculus. In this work a series of results and comparison between sequential and parallel algorithms are presented. Also, there are presented test scenarios in different cases that confirm or infirm the parallel strategy chosen. This paper will integrate the results obtained with the proposed algorithms reports or comparisons of the proposed parallelization methods and will provide conclusions and ideas for future research in this field with extensive uses.

关键词： parallel programming image processing speed-up factor processor usage

来源：评论

学校读者我要写书评

暂无评论

A Study of Internet Map Building Recognition Based on AI Technology

A Study of Internet Map Building Recognition Based on AI Tec...

引用

International conference on Computer and Communications (ICCC)

作者： Zhiheng Gao Siping Yuan Yunping Gong Emerging Information Technology Research Center Research Institute of China Telecom Co. Ltd. Guangzhou China

ISBN: (数字)9781728186351

ISBN: (纸本)9781728186368

A solution for building objects recognition from Internet map is proposed. Map pages are separately screenshot with customized CSS style and spliced together. Building contours are detected with Suzuki's contour tracing algorithm. Annotation points of text labels are detected with Hough Transform. Text labels in the image are first detected with AdvancedEAST algorithm, and then recognized as characters with CRNN algorithm, which is fine-tuned with specific data set in our works. A fast processing architecture design based on distributed parallel computing is proposed to efficiently accomplish the whole recognition process.

关键词： image recognition Text recognition Buildings Transforms parallel processing Internet Object recognition

来源：评论

学校读者我要写书评

暂无评论

parallel 3D temperature image reconstruction using multi-color magnetic particle imaging (MPI)

引用

AIP conference Proceedings 2024年第1期3230卷

作者： Klaus N. Quelhas Mark-Alexander Henn Ricardo Farias Weston L. Tew Solomon I. Woods National Institute of Standards and Technology (NIST) 100 Bureau Drive Gaithersburg MD USA Systems Engineering and Computer Science Program (PESC) Federal University of Rio de Janeiro Rio de Janeiro RJ Brazil College of Computer Mathematical & Natural Sciences University of Maryland 7998 Regents Dr College Park MD USA

Magnetic Particle Imaging (MPI) is a novel technique developed for remotely detecting magnetic nanoparticle (MNP) tracers, with great potential for biomedical imaging (as an alternative to traditional methods like MRI or CT), cell tracking, targeted drug delivery and magnetic hyperthermia. More recently, MPI has been studied as a potential method of non-contact temperature measurement. This work presents a simulation study of the multi-color MPI method tailored for 3D temperature imaging, discusses the feasibility of the method for 3D temperature measurements, and shows a parallel implementation of the multi-color T-MPI reconstruction algorithm in graphics processing unit (GPU). While the use of the parallel algorithm resulted in executions about 40x faster when compared to the serial implementation, the method exhibited serious limitations in accurately resolving particle temperatures between the calibration temperatures by interpolation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Block distributed 3MG Algorithm and its Application to 3D image Restoration

Block Distributed 3MG Algorithm and its Application to 3D Im...

引用

IEEE International conference on image processing

作者： Mathieu Chalvidal Emilie Chouzenoux Université Paris-Saclay CentraleSupélec Inria CVN Gif-sur-Yvette France

ISBN: (数字)9781728163956

ISBN: (纸本)9781728163963

Modern 3D image recovery problems require powerful optimization frameworks to handle high dimensionality while providing reliable numerical solutions in a reasonable time. In this perspective, asynchronous parallel optimization algorithms have received an increasing attention by overcoming memory limitation issues and communication bottlenecks. In this work, we propose a block distributed Majorize-Minorize Memory Gradient (BD3MG) optimization algorithm for solving large scale non-convex differentiable optimization problems. Assuming a distributed memory environment, the algorithm casts the efficient 3MG scheme into smaller dimension subproblems where blocks of variables are addressed in an asynchronous manner. Convergence of the sequence built by the proposed BD3MG method is established under mild assumptions. Application to the restoration of 3D images degraded by a depth-variant blur shows that our method yields significant computational time reduction compared to several synchronous and asynchronous competitors, while exhibiting great scalability potential.

关键词： Convergence Three-dimensional displays Optimization image restoration Delays Indexes Linear programming

来源：评论

学校读者我要写书评

暂无评论

PSO-PS:Parameter Synchronization with Particle Swarm Optimization for distributed Training of Deep Neural Networks

PSO-PS:Parameter Synchronization with Particle Swarm Optimiz...

引用

International Joint conference on Neural Networks (IJCNN)

作者： Qing Ye Yuxuan Han Yanan Sun Jiancheng Lv College of Computer Science Sichuan University Chengdu China

ISBN: (数字)9781728169262

ISBN: (纸本)9781728169279

Parameter updating is an important stage in parallelism-based distributed deep learning. Synchronous methods are widely used in distributed training the Deep Neural Networks (DNNs). To reduce the communication and synchronization overhead of synchronous methods, decreasing the synchronization frequency (e.g., every n mini-batches) is a straightforward approach. However, it often suffers from poor convergence. In this paper, we propose a new algorithm of integrating Particle Swarm Optimization (PSO) into the distributed training process of DNNs to automatically compute new parameters. In the proposed algorithm, a computing work is encoded by a particle, the weights of DNNs and the training loss are modeled by the particle attributes. At each synchronization stage, the weights are updated by PSO from the sub weights gathered from all workers, instead of averaging the weights or the gradients. To verify the performance of the proposed algorithm, the experiments are performed on two commonly used image classification benchmarks: MNIST and CIFAR10, and compared with the peer competitors at multiple different synchronization configurations. The experimental results demonstrate the competitiveness of the proposed algorithm.

关键词： Training Synchronization parallel processing Neural networks Computational modeling Mathematical model Data models

来源：评论

学校读者我要写书评

暂无评论

Scale-free distributed control for the DC MG cluster under DoS attacks: An asynchronous switched strategy

引用

Automatica 2025年 179卷

作者： Cheng, Fabin Lai, Jingang School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China Key Laboratory of Image Processing and Intelligent Control Ministry of Education Huazhong University of Science and Technology Wuhan430074 China

This paper studies the voltage restoration problem of the DC Microgrid (MG) cluster with a two-layer communication structure under Denial-of-Service (DoS) attacks. Based on the open multiagent framework, an asynchronous switching strategy is designed to suppress the impact of DoS attacks on the entire MG cluster's voltage restoration. Meanwhile, we derived the relationship between attack model parameters and communication link switching mode selection through the general model of DoS attacks and asynchronous switching methods based on average dwell time. Moreover, the designed distributed control protocol is scale-free;that is, it does not rely on any information from the two-layer communication structure and the number of distributed generations (DGs), which makes it universal. Through rigorous theoretical proof, it has been proven that in the DC MG cluster with two-layer communication structures, voltage restoration can be achieved regardless of the number and structure of DGs being attacked. Finally, two simulation examples on the MATLAB/SimPowerSystem, including a DC MG cluster consisting of 8 DGs and an IEEE 33-node system are used to investigate the scalability and flexibility of the proposed distributed control algorithm. © 2025 Elsevier Ltd

关键词： Asynchronous switched strategy DC MG cluster distributed control DoS attack Scale-free

来源：评论

学校读者我要写书评

暂无评论

Joint Channel Pruning and Quantization-Based CNN Network Learning with Mobile Computing-Based image Recognition

引用

WIRELESS COMMUNICATIONS & MOBILE COMPUTING 2021年第1期2021卷

作者： Liu, Huanyu Luo, Qing Shao, Mingmei Pan, Jeng-Shyang Li, Junbao Harbin Inst Technol Sch Elect & Informat Engn Harbin 150001 Peoples R China Nanjing Univ Aeronaut & Astronaut Coll Aerosp Engn Nanjing 210016 Peoples R China Shenyang Aircraft Design & Res Inst Shenyang 110035 Peoples R China Shandong Univ Sci & Technol Coll Comp Sci & Engn Qingdao 266590 Peoples R China

The development of the Internet and communication technology has ushered in a new era of the Internet of Things (IoT). Moreover, with the rapid development of artificial intelligence, objects are endowed with intelligence, such as home automation and smart healthcare, which are typical applications of artificial intelligence technology in IoT. With the rise of convolutional neural network (CNN) in the field of computer vision, more and more practical applications need to deploy CNN on mobile devices. However, due to the large amount of CNN computing operations and the large number of parameters, it is difficult to deploy on ordinary edge devices. The neural network model compression method has become a popular technology to reduce the computational cost and has attracted more and more attention. We specifically design a small target detection network for hardware platforms with limited computing resources, use pruning and quantization methods to compress, and demonstrate in VOC dataset and RSOD dataset on the actual hardware platform. Experiments show that the proposed method can maintain a fairly accurate rate while greatly speeding up the inference speed.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

mF2C: The Evolution of Cloud Computing Towards an Open and Coordinated Ecosystem of Fogs and Clouds 25th

mF2C: The Evolution of Cloud Computing Towards an Open and C...

引用

25th International conference on parallel and distributed Computing (Euro-Par)

作者： Masip-Bruin, Xavi Marin-Tordera, Eva Ferrer, Ana Juan Salis, Antonio Kennedy, John Jensen, Jens Jukan, Admela Bartoli, Andrea Badia, Rosa M. Cankar, Matija Begin, Marc Elian Univ Politecn Cataluna CRAAX UPC Vilanova I La Geltru Spain ATOS Res & Innovat Barcelona Spain Engn Sardegna Cagliari Italy Intel Res & Dev Leixlip Ireland UK Res & Innovat STFC RAL Didcot Oxon England Tech Univ Carolo Wilhelmina Braunschweig Braunschweig Germany Worldsensing Ltd Cambridge England Barcelona Supercomp Ctr Barcelona Spain XLAB Doo Lubjiana Slovenia SIXSQ Geneva Switzerland

ISBN: (纸本)9783030483401;9783030483395

Fog computing brings cloud computing capabilities closer to the end-devices and users, while enabling location-dependent resource allocation, low latency services, and extending significantly the IoT services portfolio as well as market and business opportunities in the cloud and IoT sectors. With the number of devices growing exponentially globally, new cloud and fog models are expected to emerge, paving the way for shared, collaborative, extensible mobile, volatile and dynamic compute, storage and network infrastructure. When put together, cloud and fog computing create a new stack of resources, which we refer to as Fog-to-Cloud (F2C), creating the need for a new, open and coordinated management ecosystem. The EU Horizon 2020 program has recently funded a new research initiative (mF2C) bringing together relevant industry and academic players in the cloud arena, aimed at designing an open, secure, decentralized, multistakeholder management framework for F2C computing, including novel programming models, privacy and security, data storage techniques, service creation, brokerage solutions, SLA policies, and resource orchestration methods. This paper introduces the main mF2C concepts, illustrates the need for a coordinated management ecosystem, proposes a preliminary design of its foundational building blocks and presents results that show the benefits mF2C may have on three key real-world scenarios.

关键词： Cloud computing Fog computing Fog-to-cloud Management distributed systems Security IoT

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：