检索结果-内蒙古大学图书馆

Cloud Deployment of High-Resolution Medical image Analysis With TOMAAT

IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS 2019年第3期23卷 969-977页

作者： Milletari, Fausto Frei, Johann Aboulatta, Moustafa Vivar, Gerome Ahmadi, Seyed-Ahmad Nvidia Inc Santa Clara CA 95051 USA German Ctr Vertigo & Balance Disorders DSGZ D-81377 Munich Germany Ludwig Maximilians Univ Munchen D-81377 Munich Germany Tech Univ Munich Garching D-85748 Garching Germany

Background: Deep learning has been recently applied to a multitude of computer vision and medical image analysis problems. Although recent research efforts have improved the state of the art, most of the methods cannot be easily accessed, compared or used by other researchers or clinicians. Even if developers publish their code and pre-trained models on the internet, integration in stand-alone applications and existing workflows is often not straightforward, especially for clinical research partners. In this paper, we propose an open-source framework to provide AI-enabled medical image analysis through the network. methods: TOMAAT provides a cloud environment for general medical image analysis, composed of three basic components: (i) an announcement service, maintaining a public registry of (ii) multiple distributed server nodes offering various medical image analysis solutions, and (iii) client software offering simple interfaces for users. Deployment is realized through HTTP-based communication, along with an API and wrappers for common image manipulations during pre- and post-processing. Results: We demonstrate the utility and versatility of TOMAAT on several hallmark medical image analysis tasks: segmentation, diffeomorphic deformable atlas registration, landmark localization, and workflow integration. Through TOMAAT, the high hardware demands, setup and model complexity of demonstrated approaches are transparent to users, who are provided with simple client interfaces. We present example clients in three-dimensional Slicer, in the web browser, on iOS devices and in a commercially available, certified medical image analysis suite. Conclusion: TOMAAT enables deployment of state-of-the-art image segmentation in the cloud, fostering interaction among deep learning researchers and medical collaborators in the clinic. Currently, a public announcement service is hosted by the authors, and several ready-to-use services are registered and enlisted at http://***

关键词： Deep learning medical image analysis segmentation registration cloud deployment clinical translation

来源：评论

学校读者我要写书评

暂无评论

PowerSGD: Practical Low-Rank Gradient Compression for distributed Optimization 33

PowerSGD: Practical Low-Rank Gradient Compression for Distri...

引用

33rd Conference on Neural Information processing Systems (NeurIPS)

作者： Vogels, Thijs Karinireddy, Sai Praneeth Jaggi, Martin Ecole Polytech Fed Lausanne Lausanne Switzerland

We study lossy gradient compression methods to alleviate the communication bottleneck in data-parallel distributed optimization. Despite the significant attention received, current compression schemes either do not scale well, or fail to achieve the target test accuracy. We propose a new low-rank gradient compressor based on power iteration that can i) compress gradients rapidly, ii) efficiently aggregate the compressed gradients using all-reduce, and iii) achieve test performance on par with SGD. The proposed algorithm is the only method evaluated that achieves consistent wall-clock speedups when benchmarked against regular SGD using highly optimized off-the-shelf tools for distributed communication. We demonstrate reduced training times for convolutional networks as well as LSTMs on common datasets. Our code is available at https://***/epfml/powersgd.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Supervised Multi-View distributed Hashing

Supervised Multi-View Distributed Hashing

引用

IEEE International Conference on image processing

作者： Yunpeng Tang Xiaobo Shen Zexuan Ji Tao Wang Peng Fu Quan-Sen Sun School of Computer Science and Engineering Nanjing University of Science and Technology China

ISBN: (数字)9781728163956

ISBN: (纸本)9781728163963

Multi-view hashing efficiently integrates multi-view data for learning compact hash codes, and achieves impressive large-scale retrieval performance. In real-world applications, multi-view data are often stored or collected in different locations, where hash code learning is more challenging yet less studied. To fulfill this gap, this paper proposes a novel supervised multi-view distributed hashing (SMvDisH) for hash code learning from multi-view data in a distributed manner. SMvDisH yields the discriminative latent hash codes by joint learning of latent factor model and classifier. With local consistency assumption among neighbor nodes, the distributed learning problem is divided into a set of decentralized sub-problems. The sub-problems can be solved in parallel, and the computational and communication costs are low. Experimental results on three large-scale image datasets demonstrate that SMvDisH achieves competitive retrieval performance and trains faster than state-of-the-art multi-view hashing methods.

关键词： distributed databases Convex functions Benchmark testing Linear programming Computational modeling Binary codes Optimization

来源：评论

学校读者我要写书评

暂无评论

Classification Of Human Facial Expressions For Emotion Recognition Using A distributed Computer System

Classification Of Human Facial Expressions For Emotion Recog...

引用

International Conference on Computer Science and Engineering (UBMK)

作者： Timuçin Korkmaz Hamza Erol Department of Computer Engineering Mersin University Mersin Turkey

ISBN: (数字)9781728175652

ISBN: (纸本)9781728175669

A computer system was developed to classify human facial expressions for emotion recognition using a distributed computer architecture within the scope of big data. Computers with normal standards and features were used in the distributed computer architecture. Necessary system software for the distributed computer architecture was established and mutual communication protocols were provided between the computers and databases in the computer network. Visual C# parallel programming was used as the software language on the distributed computer architecture. In the software prepared, face image flies were processed and threads were created. The threads created later were processed in the processors of the computers. The threads were run on the processors in the distributed computer system, and facial expressions were classified for emotion recognition. In the distributed computer architecture: the number of image files, the volume of big data and the load of the threads to be processed are taken into account; and databases and parallel programming were used and the classification of human face images was performed. Emotion analysis methods and facial expression recognition techniques were used in the classification process. The distributed computer system is in low cost and high processing speed. With the distributed computer system architecture, big data analytics has become convenient and feasible.

关键词： Face recognition distributed databases Computer architecture IP networks Servers Emotion recognition

来源：评论

学校读者我要写书评

暂无评论

Accelerating Deep Learning Tasks with Optimized GPU-assisted image Decoding

Accelerating Deep Learning Tasks with Optimized GPU-assisted...

引用

International Conference on parallel and distributed Systems (ICPADS)

作者： Lipeng Wang Qiong Luo Shengen Yan HKUST Hong Kong China SenseTime Research Shenzhen China

ISBN: (数字)9781728190747

ISBN: (纸本)9781728183824

In computer vision deep learning (DL) tasks, most of the input image datasets are stored in the JPEG format. These JPEG datasets need to be decoded before DL tasks are performed on them. We observe two problems in the current JPEG decoding procedures for DL tasks: (1) the decoding of image entropy data in the decoder is performed sequentially, and this sequential decoding repeats with the DL iterations, which takes significant time; (2) Current parallel decoding methods under-utilize the massive hardware threads on GPUs. To reduce the image decoding time, we introduce a pre-scan mechanism to avoid the repeated image scanning in DL tasks. Our pre-scan generates boundary markers for entropy data so that the decoding can be performed in parallel. To cooperate with the existing dataset storage and caching systems, we propose two modes of the pre-scan mechanism: a compatible mode and a fast mode. The compatible mode does not change the image file structure so pre-scanned files can be stored back to disk for subsequent DL tasks. In comparison, the fast mode crafts a JPEG image into a binary format suitable for parallel decoding, which can be processed directly on the GPU. Since the GPU has thousands of hardware threads, we propose a fine-grained parallel decoding method on the pre-scanned dataset. The fine-grained parallelism utilizes the GPU effectively, and achieves speedups of around 1.5× over existing GPU-assisted image decoding libraries on real-world DL tasks.

关键词： Deep learning Transform coding Graphics processing units Hardware Decoding Task analysis Iterative decoding

来源：评论

学校读者我要写书评

暂无评论

image segmentation with local active contours on graphics processing units. 19

Image segmentation with local active contours on graphics pr...

引用

4th International Conference on Smart City Applications (SCA)

作者： Ammar, Abderazzak Bouattane, Omar Youssfi, Mohammed Univ Hassan II Casablanca Lab SSDIA ENSET Mohammadia Morocco

ISBN: (纸本)9781450362894

Active contours model (ACM) is one of the most attractive and popular methods for image segmentation. While the different formulations of the ACM are generally known to be highly computationally demanding, like most of the image processing techniques do, they are however potential candidates for parallelization. The graphics processing units (GPUs) are nowadays increasingly present in all kind of image processing among other research areas. In this paper, our goal was the implementation of a GPU parallelized local based formulation of the active contours model, namely the local binary fitting energy (LBF) algorithm. This algorithm involves a large amount of convolution operation computations. Besides the convolution, many other finite differences based operations involved in the algorithm are readily parallelized. So, almost all the heavy computational work is moved to the GPU side, and the CPU needs only to control the main loop and some intermediate computations. To validate our results, we have been experimenting with medical images of different sizes, and different convolution filtering sizes, by comparing the speedup of the GPU parallel implementation vs the CPU sequential implementation.

关键词： active contours model segmentation parallelization graphics processing units GPU medical images

来源：评论

学校读者我要写书评

暂无评论

Handbook of Automata Theory

引用

2021年

作者： Jean-Éric Pin

ISBN: (数字)9783985475063

ISBN: (纸本)9783985470068

Automata theory is a subject of study at the crossroads of mathematics, theoretical computer science, and applications. In its core it deals with abstract models of systems whose behaviour is based on transitions between states, and it develops methods for the description, classification, analysis, and design of such systems. The Handbook of Automata Theory gives a comprehensive overview of current research in automata theory, and is aimed at a broad readership of researchers and graduate students in mathematics and computer science. Volume I is divided into three parts. The first part presents various types of automata: automata on words, on infinite words, on finite and infinite trees, weighted and maxplus automata, transducers, and two-dimensional models. Complexity aspects are discussed in the second part. Algebraic and topological aspects of automata theory are covered in the third part. Volume ii consists of two parts. The first part is dedicated to applications of automata in mathematics: group theory, number theory, symbolic dynamics, logic, and real functions. The second part presents a series of further applications of automata theory such as message-passing systems, symbolic methods, synthesis, timed automata, verification of higher-order programs, analysis of probabilistic processes, natural language processing, formal verification of programs and quantum computing. The two volumes comprise a total of thirty-nine chapters, with extensive references and individual tables of contents for each one, as well as a detailed subject index.

关键词： finite automata Hopcroft's algorithm regular languages regular expressions finite transducers weighted automata automata on infinite words automata on trees picture languages algorithmic learning descriptional complexity Boolean circuits synchronizing automata road colouring problem varieties of languages profinite topology descriptive set theory equational theories language equations forest algebras free groups Stallings automata automatic groups self-similar groups automatic sequences numeration systems Cobham's theorem symbolic dynamics synchronous automata finite model theory fractal image generation communicating automata model-checking Church's problem distributed synthesis timed automata recursion schemes Markov decision processes infinite-state systems natural language processing temporal logic branching time logic quantum finite automata

来源：评论

学校读者我要写书评

暂无评论

船舶航行交通事件实时检测研究现状与展望

引用

交通信息与安全 2022年

作者：黄琛陈德山严新平吴兵武汉理工大学交通与物流工程学院武汉理工大学国家水运安全工程技术研究中心武汉理工大学智能交通系统研究中心

船舶航行交通事件检测多采取基于历史数据的离线检测方式，检测模型适用性差，无法满足监管人员实时检测需求。通过对现有船舶异常行为和船舶事故检测文献的分析，可以发现：在数据层面，数据来源单一、环境信息缺失；在方法层面，基于... 详细信息

船舶航行交通事件检测多采取基于历史数据的离线检测方式，检测模型适用性差，无法满足监管人员实时检测需求。通过对现有船舶异常行为和船舶事故检测文献的分析，可以发现：在数据层面，数据来源单一、环境信息缺失；在方法层面，基于统计、风险评估的经典模型检测效率高但检测准确性低，基于神经网络、图像识别等机器学习的检测方法准确性高但效率低；多源数据融合、多项技术结合的检测方法成为实时检测趋势。梳理船舶航行交通事件实时检测关键技术3项。①海事大数据处理技术：高效处理船舶运动数据和航行环境数据，标准化多源异构数据结构，降低数据源单一造成的事件报警频率，扩大数据适用性；②船舶行为动态建模技术：利用深度学习、语义关联、图神经网络等方法融合船舶运动数据与船舶航行情境信息，考虑船舶交通的空间相关性及环境影响，提高检测准确性和实时性；③实时分析和可视化技术：结合平行系统进行虚实系统间信息传递，定性分析检测结果，实时显示检测全过程，提升监管交互性。面向海事监管未来的发展方向，结合新一代航运系统，提出平行检测系统架构，包括数据采集、后台服务和客户端应用3个功能模块，实时接收处理船舶航行数据、分析预测交通状态、动态检测预警交通事件并展示仿真结果；结合平行检测系统和实时检测关键技术，从数据融合处理、交通状态感知和交通虚实映射3个方面进行展望，提出实际使用阶段的实时检测技术发展方向。

关键词：水路交通船舶航行交通事件实时检测平行检测系统异常行为船舶事故

来源：评论

学校读者我要写书评

暂无评论

Geometry-based solution of joint diagonalisation in blind source separation

引用

IET SIGNAL processing 2019年第9期13卷 755-765页

作者： Akhavan, Saeed Esmaeili, Samaneh Soltanian-Zadeh, Hamid Univ Tehran Sch Elect & Comp Engn Tehran Iran Univ Grenoble Alpes CNRS Grenoble INP GIPSA Lab Grenoble France Henry Ford Hlth Syst Med Image Anal Lab Detroit MI 48202 USA

Joint diagonalisation (JD) of a set of target matrices is a common approach to solve the blind source separation (BSS) problem. In fact, the separating matrix (the inverse of the mixing matrix) of the sources is the joint diagonaliser of the target matrices. In this study, the authors show that each row of the separating matrix maps the target matrices using a linear mapping into the new vectors which are located on a direct line along the corresponding column of the mixing matrix. Based on this geometrical interpretation, the authors propose a method for solving JD problem, which let us estimate the rows of the separating matrix and the columns of the mixing matrix, (i) independently and in parallel, (ii) consecutively or (iii) simultaneously. Simulation results in different scenarios such as additive noise, ill-posed mixing matrix, and high-dimensional data demonstrate the effectiveness of the proposed method relative to state-of-the-art JD methods. The authors also propose an approach to omit the effect of outlier data, which severely degrades parameters estimation. The proposed approach can be used in JD algorithms to improve their performance in the presence of outlier data.

关键词： blind source separation parameter estimation iterative methods matrix algebra geometry geometry-based solution joint diagonalisation target matrices blind source separation problem joint diagonaliser separating matrix maps linear mapping JD problem ill-posed mixing matrix state-of-the-art JD methods

来源：评论

学校读者我要写书评

暂无评论

DSPG: Decentralized Simultaneous Perturbations Gradient Descent Scheme

DSPG: Decentralized Simultaneous Perturbations Gradient Desc...

引用

Euromicro Conference on parallel, distributed and Network-Based processing

作者： Arunselvan Ramaswamy Dept. of Electrical and Computer Science Engineering Paderborn University Paderborn Germany

ISBN: (数字)9781728165820

ISBN: (纸本)9781728165837

distributed descent-based methods are an essential toolset to solving optimization problems in multi-agent system scenarios. Here the agents seek to optimize a global objective function through mutual cooperation. Oftentimes, cooperation is achieved over a wireless communication network that is prone to delays and errors. There are many scenarios wherein the objective function is either non-differentiable or merely observable. In this paper, we present a cross-entropy based distributed stochastic approximation algorithm (SA) that finds a minimum of the objective, using only samples. We call this algorithm Decentralized Simultaneous Perturbation Stochastic Gradient, with Constant Sensitivity Parameters (DSPG). This algorithm is a two fold improvement over the classic Simultaneous Perturbation Stochastic Approximations (SPSA) algorithm. Specifically, DSPG allows for (i) the use of old information from other agents and (ii) easy implementation through the use simple hyper-parameter choices. We analyze the biases and variances that arise due to these two allowances. We show that the biases due to communication delays can be countered by a careful choice of algorithm hyper-parameters. The variance of the gradient estimator and its effect on the rate of convergence is studied. We present numerical results supporting our theory. Finally, we discuss an application to the stochastic consensus problem.

关键词： Sensitivity Approximation algorithms Delays Linear programming Convergence Gradient methods

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：