检索结果-内蒙古大学图书馆

International symposium on Computing and Networking (CANDAR)

作者： Yuttakon Yuttakonkit Yasuhiko Nakashima Graduate School of Information Science Nara Institute of Science and Technology Japan

ISBN: (纸本)9781509026562

Recently, many approaches apply light-field image processing on smartphones and wearable devices. A Graphic processing Unit (GPU) is commonly used to exploit parallelism in such image processing. However, because the access pattern in the light-field application is more sparse than typical stencil applications and does not use all data in a cache line. Furthermore, the data requests to multiple locations generate enormous short-burst memory transfers in the cache system, cost high latency, and do not fully utilize the high memory bandwidth of GPU. Therefore, an alternative architecture that exploits a long-burst data transmission, which improves the memory bandwidth utilization, is essential. We propose a sparse stencil oriented Coarse Grain Reconfigurable Accelerator (CGRA) that we call EMAXV. Unlike on-demand multiple data loading on GPU, EMAXV loads the input data with a long burst transferring before the execution proceeds to conceal the sparse memory access and multi-threading cache races. It further obscures the memory loading latency with an execution latency from different activations. We evaluated the EMAXV and mobile GPU (Tegra K1) performances with identical host CPU's frequency and main memory bandwidth. Although EMAXV has much lower computation capability, we achieved four times performance of mobile GPU for light-field depth extraction and 89% of the performance for light-field image rendering.

关键词： graphics processing units image processing Instruction sets Mobile communication Bandwidth computer architecture Pipelines

来源：评论

学校读者我要写书评

暂无评论

Immersive Molecular Visualization with Omnidirectional Stereoscopic Ray Tracing and Remote Rendering

Immersive Molecular Visualization with Omnidirectional Stere...

引用

IEEE International symposium on Parallel and Distributed processing Workshops and Phd Forum (IPDPSW)

作者： John E. Stone William R. Sherman Klaus Schulten Beckman Institute University of Illinois at Urbana-Champaign Urbana IL USA Pervasive Technology Institute Indiana University Bloomington IN USA Department of Physics University of Illinois at Urbana-Champaign Urbana IL USA

ISBN: (纸本)9781509036837

Immersive molecular visualization provides the viewer with intuitive perception of complex structures and spatial relationships that are of critical interest to structural biologists. The recent availability of commodity head mounted displays (HMDs) provides a compelling opportunity for widespread adoption of immersive visualization by molecular scientists, but HMDs pose additional challenges due to the need for low-latency, high-frame-rate rendering. State-of-the-art molecular dynamics simulations produce terabytes of data that can be impractical to transfer from remote supercomputers, necessitating routine use of remote visualization. Hardware-accelerated video encoding has profoundly increased frame rates and image resolution for remote visualization, however round-trip network latencies would cause simulator sickness when using HMDs. We present a novel two-phase rendering approach that overcomes network latencies with the combination of omnidirectional stereoscopic progressive ray tracing and high performance rasterization, and its implementation within VMD, a widely used molecular visualization and analysis tool. The new rendering approach enables immersive molecular visualization with rendering techniques such as shadows, ambient occlusion lighting, depth-of-field, and high quality transparency, that are particularly helpful for the study of large biomolecular complexes. We describe ray tracing algorithms that are used to optimizeinteractivity and quality, and we report key performance metrics of the system. The new techniques can also benefit many other application domains.

关键词： Rendering (computer graphics) Stereo image processing Ray tracing Visualization Lighting graphics processing units Head

来源：评论

学校读者我要写书评

暂无评论

Steps in deployment and development of Convolutional Neural Network based applications

Steps in deployment and development of Convolutional Neural ...

引用

International symposium on Electronics and Telecommunications (ISETC)

作者： Şerban Marcel Măduţa Cătălin Daniel Căleanu Faculty of Electronics and Telecommunications POLITEHNICA University Timişoara Timişoara România

Designated among 10 breakthrough technologies by MIT Technology Review [1], Deep Learning (DL) outperform current approaches in many situations, e.g. image or speech processing. One of the most important deep architecture is represented by the Convolutional Neural Network (CNN). The purpose of this paper is to provide practical recommendations in the deployment and development of the CNN based applications. They refer to the hardware as well as software available solutions and go beyond by providing guidance in choosing the appropriate hyper-parameters (structure, training algorithm, learning rate, regularization techniques, etc.). The experimental results are reported using the CIFAR-10 dataset.

关键词： Training Neural networks Machine learning computer architecture graphics processing units Hardware

来源：评论

学校读者我要写书评

暂无评论

Reducing Visual Discomfort with HMDs Using Dynamic Depth of Field

引用

IEEE computer graphics AND APPLICATIONS 2015年第5期35卷 34-41页

作者： Carnegie, Kieran Rhee, Taehyun Victoria Univ Wellington Sch Engn & Comp Sci Wellington New Zealand

Although head-mounted displays (HMDs) are ideal devices for personal viewing of immersive stereoscopic content, exposure to VR applications on them results in significant discomfort for the majority of people, with symptoms including eye fatigue, headaches, nausea, and sweating. A conflict between accommodation and vergence depth cues on stereoscopic displays is a significant cause of visual discomfort. This article describes the results of an evaluation used to judge the effectiveness of dynamic depth-of-field (DoF) blur in an effort to reduce discomfort caused by exposure to stereoscopic content on HMDs. Using a commercial game engine implementation, study participants report a reduction of visual discomfort on a simulator sickness questionnaire when DoF blurring is enabled. The study participants reported a decrease in symptom severity caused by HMD exposure, indicating that dynamic DoF can effectively reduce visual discomfort.

关键词： Visualization Stereo image processing Hardware Virtual Reality Software Visual Systems Displays Virtual Reality computer graphics Immersive Stereoscopic Content Depth Of Field Head Mounted Display

来源：评论

学校读者我要写书评

暂无评论

Development of a Software for Coaching of Paralympic Sport Shooting 12

Development of a Software for Coaching of Paralympic Sport S...

引用

12th Latin American Robotics symposium / 3rd brazilian symposium on Robotics (LARS-SBR)

作者： Franco de Morais, Renan Alves Rezende, Arthur Ribeiro Goncalves, Rogerio Sales Univ Fed Uberlandia Sch Mech Engn Uberlandia MG Brazil

ISBN: (纸本)9781467371292

This study proposes the development of a virtual environment, as a computer game format, that aid the training and the practice of Paralympic sport shooting, using a digital image analysis system, involving techniques of acquisition and processing of acquired images. The image acquisition was performed by a webcam that interacts with the software, which performs the processing and image analysis in matrix form, displaying the results dynamically inside the game's screen. It can be said as a conclusion that it was possible to use programming techniques in Matlab, combined with the capture and image processing, to create a training tool able to simulate the practice of Paralympic sport shooting, providing the appropriate interactivity using graphic interfaces.

关键词： image analysis webcam Matlab sport shooting paralympic

来源：评论

学校读者我要写书评

暂无评论

Unconditionally Stable Shock Filters for image and Geometry processing 15

Unconditionally Stable Shock Filters for Image and Geometry ...

引用

symposium on Geometry processing

作者： Prada, F. Kazhdan, M. Johns Hopkins Univ Baltimore MD 21218 USA

This work revisits the Shock Filters of Osher and Rudin [OR90] and shows how the proposed filtering process can be interpreted as the advection of image values along flow-lines. Using this interpretation, we obtain an efficient implementation that only requires tracing flow-lines and re-sampling the image. We show that the approach is stable, allowing the use of arbitrarily large time steps without requiring a linear solve. Furthermore, we demonstrate the robustness of the approach by extending it to the processing of signals on meshes in 3D.

关键词： Categories and Subject Descriptors (according to ACM CCS) I.3.5 [computer graphics]: Computational Geometry and Object Modeling - Hierarchy and geometric transformations I.4.3 [image processing and computer Vision]: Enhancement - Sharpening and deblurring

来源：评论

学校读者我要写书评

暂无评论

Progressive feedback point cloud rendering for virtual reality display

Progressive feedback point cloud rendering for virtual reali...

引用

IEEE Annual International symposium Virtual Reality

作者： Ross Tredinnick Markus Broecker Kevin Ponto Wisconsin Institute for Discovery - University of Wisconsin-Madison

ISBN: (纸本)9781509008377

Previous approaches to rendering large point clouds on immersive displays have generally created a trade-off between interactivity and quality. While these approaches have been quite successful for desktop environments when interaction is limited, virtual reality systems are continuously interactive, which forces users to suffer through either low frame rates or low image quality. This paper presents a novel approach to this problem through a progressive feedback-driven rendering algorithm. This algorithm uses reprojections of past views to accelerate the reconstruction of the current view. The presented method is tested against previous methods, showing improvements in both rendering quality and interactivity.

关键词： Rendering (computer graphics) Three-dimensional displays graphics processing units Measurement Virtual reality Feedback loop Geometry

来源：评论

学校读者我要写书评

暂无评论

Sensing-by-Overlaying: A Practical Implementation of a Multiplayer Mixed-Reality Gaming System by Integrating a Dense Point Cloud and a Real-Time Camera

Sensing-by-Overlaying: A Practical Implementation of a Multi...

引用

IEEE International symposium on Multimedia

作者： Shuichi Kurabayashi Hidetoshi Ishiyama Masaru Kanai Cygames Research Cygames Inc. Tokyo Japan

ISBN: (纸本)9781509045723

This paper presents a mixed-reality gaming system architecture with a two-phase spatial information processing model. This model assumes that the physical world consists of mainly stable elements such as buildings, and wide-area physical structures are captured in advance by using a 3D laser range scanner. The captured structure data has exactly the same structure and topology as the physical world. Our system recognizes the physical world at runtime by overlaying this virtual structure data on a runtime monocular camera image at exactly the same position. By adopting this model, it is straightforward to implement image-based lighting for photometric integrity and occlusion handling for geometric integrity, because the overlaid virtual structure provides precise distance metrics and depth information. This model also resolves conflicts involving multiple sensing results obtained by multiple users. This paper shows the results of empirical evaluation experiments using the implemented prototype system.

关键词： Three-dimensional displays Sensors Lighting Games Rendering (computer graphics) Virtual reality Runtime

来源：评论

学校读者我要写书评

暂无评论

Binary image classification using a neurosynaptic processor: A trade-off analysis

Binary image classification using a neurosynaptic processor:...

引用

IEEE International symposium on Circuits and Systems (ISCAS)

作者： William Murphy Megan Renz Qing Wu Air Force Research Lab University at Buffalo

ISBN: (纸本)9781479953424

This paper examines the performance of two power efficient hardware implementations using deep neural networks to perform a simple image classification task. We provide the first ever examination of the accuracy-energy trade-offs of deep neural networks running on both an embedded GPU, and a neuromorphic processor. IBM's TrueNorth is a brain-inspired event-driven neuromorphic processor. It was designed to be scalable and to consume extremely low amounts of power. NVIDIA's Tegra K1 SoC is a mobile processor also designed with low power and a small footprint in mind. While these two chips were designed with similar constraints, the resulting architectures and performance trade-offs achieved are significantly different. On our simple image classification task Convolutional Neural Networks utilizing the Tegra K1 SoC achieve up to 89 % accuracy with a normalized accuracy per active energy, ||Acc||/EA, score of up to 24.22 on our test dataset, while Tea Networks running on the TrueNorth processor achieve less accuracy at 82%, but a better accuracy-energy trade-off with a ||Acc||/EA score of up to 158.49.

关键词： Neurons Training graphics processing units computer architecture Measurement Sensors Algorithm design and analysis

来源：评论

学校读者我要写书评

暂无评论

UFO — a scalable platform for high-speed synchrotron X-ray imaging

UFO — a scalable platform for high-speed synchrotron X-ray ...

引用

IEEE symposium on Nuclear Science (NSS/MIC)

作者： Andreas Kopmann Suren Chilingaryan Matthias Vogelgesang Timo Dritschler Andrey Shkarin Roman Shkarin Tomy dos Santos Rolo Tomas Farago Thomas van de Kamp Matthias Balzer Michele Caselle Marc Weber Tilo Baumbach Karlsruhe Institute of Technology Germany

New imaging stations aim for high spatial and temporal resolution and are characterized by ever increasing sampling rates and demanding data processing workflows. Key to successful imaging experiments is to open up high-performance computing resources. This includes carefully selected components for computing hardware and development of advanced imaging algorithms optimized for efficient use of parallel processor architectures. We present the novel UFO computing platform for online data processing for imaging experiments and image-based feedback. The platform handles the full data life cycle from the X-ray detector to long-term data archives. Core components of this system are an FPGA platform for ultra-fast data acquisition, the GPU-based UFO image processing framework, and the fast control system “Concert”. Reconstruction algorithms implemented in the UFO framework are optimized for the latest GPU architectures and provide a reconstruction throughput in the GB/s-range. The control system “Concert” integrates high-speed computing nodes and fast beamline devices and thus enables image-based control loops and advanced workflow automation for efficient beam time usage. Low latencies are ensured by direct communication between FPGA and GPUs using AMDs DirectGMA technology. Time resolved tomography is supported by cutting edge regularization methods for high quality reconstructions with a reduced number of projections. The new infrastructure at ANKA has dramatically accelerated tomography from hours to second and resulted in new application fields, like high-throughput tomography, pump-probe radiography and stroboscopic tomography. Ultra-fast X-ray cine-tomography for the first time allows one to observe internal dynamics of moving millimeter-sized objects in real-time.

关键词： Cameras Tomography image reconstruction Control systems graphics processing units computer architecture

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：