检索结果-内蒙古大学图书馆

A Parallel Distributed processing Algorithm for image Feature Extraction 1

14th International symposium on Intelligent Data Analysis (IDA)

作者： Belousov, Alexander Ratsaby, Joel Ariel Univ Elect & Elect Engn Dept Ariel Israel

ISBN: (数字)9783319244655

ISBN: (纸本)9783319244655

We present a new parallel algorithm for image feature extraction. which uses a distance function based on the LZ-complexity of the string representation of the two images. An input image is represented by a feature vector whose components are the distance values between its parts (sub-images) and a set of prototypes. The algorithm is highly scalable and computes these values in parallel. It is implemented on a massively parallel graphics processing unit (GPU) with several thousands of cores which yields a three order of magnitude reduction in time for processing the images. Given a corpus of input images the algorithm produces labeled cases that can be used by any supervised or unsupervised learning algorithm to learn image classification or image clustering. A main advantage is the lack of need for any image processing or image analysis;the user only once defines image-features through a simple basic process of choosing a few small images that serve as prototypes. Results for several image classification problems are presented.

关键词： image classification

来源：评论

学校读者我要写书评

暂无评论

A Performance Model and Efficiency-Based Assignment of Buffering Strategies for Automatic GPU Stencil Code Generation

A Performance Model and Efficiency-Based Assignment of Buffe...

引用

IEEE International symposium on Embedded Multicore Socs (MCSoC)

作者： Yue Hu David M. Koppelman Steven R. Brandt Division of Electrical & Computer Engineering Louisiana State University Baton Rouge LA USA Division of Computer Science Louisiana State University Baton Rouge LA USA

ISBN: (纸本)9781509035328

Stencil computations form the basis for computer simulations across almost every field of science, such as computational fluid dynamics, data mining, and image processing. Their mostly regular data access patterns potentially enable them to take advantage of the high computation and data bandwidth of GPUs, but only if data buffering and other issues are handled properly. Finding a good code generation presents a number of challenges, one of which is the best way to make use of memory. GPUs have three types of on-chip storage: registers, shared memory, and read-only cache. The choice of type of storage and how it's used, a buffering strategy, for each stencil array (grid function, (GF)) not only requires a good understanding of its stencil pattern, but also the efficiency of each type of storage for the GF, to avoid squandering storage that would be more beneficial to another GF. Our code-generation framework supports five buffering strategies. For a stencil computation with N GFs, the total number of possible assignments is 5N. Large, complex stencil kernels may consist of dozens of GFs, resulting in significant search overhead. In this work, we present an analytic performance model for stencil computations on GPUs, and study the behavior of readonly cache and L2 cache. Next, we propose an efficiency-based assignment algorithm, which operates by scoring a change in buffering strategy for a GF using a combination of (a) the predicted execution time and (b) on-chip storage usage. By using this scoring an assignment for N GFs and b strategy types can be determined in (b - 1)N(N + 1)/2 steps. Results show that the performance model has good accuracy and that the assignment strategy is highly efficient.

关键词： Registers Computational modeling Instruction sets graphics processing units Kernel System-on-chip Buffer storage

来源：评论

学校读者我要写书评

暂无评论

Nonlinear waveform inversion by use of the regularized dual averaging method for ultrasound computed tomography

Nonlinear waveform inversion by use of the regularized dual ...

引用

Progress in Electromagnetic Research symposium (PIERS)

作者： Thomas P. Matthews Kun Wang Cuiping Li Neb Duric Mark A. Anastasio Washington University in Saint Louis Saint Louis MO US Delphinus Medical Technologies USA

ISBN: (纸本)9781509060948

Ultrasound computed tomography (USCT) is a promising medical imaging modality that offers several novel tissue contrasts and holds great potential for breast cancer screening. Waveform inversion-based image reconstruction methods account for higher order diffraction effects and can produce high-resolution USCT images, but are computationally demanding. Recently, a source encoding technique was combined with stochastic gradient descent (SGD) to greatly reduce image reconstruction times. However, this method bundles the stochastic data fidelity term with the deterministic regularization term. This non-optimal approach can be overcome by replacing stochastic gradient descent with a structured optimization method, such as the regularized dual averaging (RDA) method, that exploits knowledge of the composition of the cost function. In this work, the dual averaging method is combined with source encoding techniques to improve the effectiveness of regularization while maintaining the reduced reconstruction times afforded by source encoding. It is demonstrated that, for a particular choice of regularization function, each iteration can be decomposed into a gradient descent step based on the data fidelity term and a proximal update step corresponding to the regularization term. Furthermore, the regularization term is never explicitly differentiated, allowing non-smooth regularization penalties to be naturally incorporated. The wave equation is solved by use of a time-domain k-space pseudo-spectral method, implemented on general-purpose graphics processing units. The effectiveness of this approach is demonstrated through computer simulation and experimental studies. The results suggest that the dual averaging method can produce images with less noise and comparable resolution to those obtained by use of stochastic gradient descent.

关键词： Ultrasonic imaging Computed tomography Biomedical imaging image reconstruction Encoding Electromagnetics Electromagnetic scattering

来源：评论

学校读者我要写书评

暂无评论

The Use of a Graphic processing Unit (GPU) in a Real Time Visual Odometry Application 45

The Use of a Graphic Processing Unit (GPU) in a Real Time Vi...

引用

45th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops, DSN-W 2015

作者： Vargas, Jaime Armando Delgado Kurka, Paulo Roberto Gardel Faculdade de Engenharia Mecânica Universidade Estadual de Campinas CampinasSão Paulo Brazil

ISBN: (纸本)9781467380447

This paper presents a practical application of visual odometry (VO). Visual odometry applications are computationally expensive due to the frequent and large number of required data processing. In the present work the application is implemented in a graphics processing unit card (GPU) using compute unified device architecture CUDA and OpenCV libraries, allowing real time processing with a speed of 30 frames per second. The algorithm begins with the capture and processing of stereoscopic images to find invariant interest points (keypoints) using the GPU-OpenCV speed-up robust features (SURF) library implementation. Stereoscopic image points are projected in the Euclidean space to yield 3-D estimates of the robot's translation and rotation movements. The real time VO algorithm is applied in a practical odometry estimation in a robot's outdoors navigation experiment. © 2015 IEEE.

关键词： graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Cambricon-X: An accelerator for sparse neural networks 49

Cambricon-X: An accelerator for sparse neural networks

引用

IEEE/ACM International symposium on Microarchitecture (MICRO)

作者： Shijin Zhang Zidong Du Lei Zhang Huiying Lan Shaoli Liu Ling Li Qi Guo Tianshi Chen Yunji Chen Cambricon Technologies Co. Ltd. China University of Chinese Academy of Sciences Beijing China SKL of Computer Architecture CAS Beijing China Institute of Automation CAS Beijing China

ISBN: (纸本)9781509035090

Neural networks (NNs) have been demonstrated to be useful in a broad range of applications such as image recognition, automatic translation and advertisement recommendation. State-of-the-art NNs are known to be both computationally and memory intensive, due to the ever-increasing deep structure, i.e., multiple layers with massive neurons and connections (i.e., synapses). Sparse neural networks have emerged as an effective solution to reduce the amount of computation and memory required. Though existing NN accelerators are able to efficiently process dense and regular networks, they cannot benefit from the reduction of synaptic weights. In this paper, we propose a novel accelerator, Cambricon-X, to exploit the sparsity and irregularity of NN models for increased efficiency. The proposed accelerator features a PE-based architecture consisting of multiple processing Elements (PE). An Indexing Module (IM) efficiently selects and transfers needed neurons to connected PEs with reduced bandwidth requirement, while each PE stores irregular and compressed synapses for local computation in an asynchronous fashion. With 16 PEs, our accelerator is able to achieve at most 544 GOP/s in a small form factor (6.38 mm 2 and 954 mW at 65 nm). Experimental results over a number of representative sparse networks show that our accelerator achieves, on average, 7.23x speedup and 6.43x energy saving against the state-of-the-art NN accelerator.

关键词： Artificial neural networks Neurons Biological neural networks computer architecture graphics processing units Architecture Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Detector modeling in PET list-mode reconstruction: comparison between pre-calculated and on-the-fly computed system matrix

Detector modeling in PET list-mode reconstruction: compariso...

引用

IEEE Nuclear Science symposium and Medical Imaging Conference (NSS/MIC)

作者： Autret, Awen Moreau, Matthieu Carlier, Thomas Bert, Julien Strauss, Olivier Visvikis, Dimitris CHRU Brest LaTIM INSERM UMR1101 Brest France INSERM UMR S 892 Nantes France Univ Montpellier 2 CNRS UMR5506 LIRMM Montpellier France

ISBN: (纸本)9781467398626

In List-mode PET iterative image reconstruction two major approaches exist in order to model the detector response. The detector response can be measured or estimated before the reconstruction and stored in a matrix. This approach has the advantage of being highly faithful with the real detector response but requires a large among of memory to store the huge system response matrix (SRM). The other approach is based on an on-the-fly projector to compute the required parts of the SRM. While existing for multiple decades, the projector approach was limited to over simplistic detector models because of its computation burden. Since the growth of the general-purpose processing on graphics processing units that allows a large computing power in a simple computer, more complex on-the-fly projectors are now possible. We compared the first approach with the on-the-fly approach in terms of image quantitative and qualitative performances as well as reconstruction speed. Results showed that a simplistic on-the-fly projector cannot reach the image quality of a stored SRM but a more accurate on-the-fly projector can surpass the SRM in term of spatial resolution while running faster especially because a GPU implementation can be used.

关键词： graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

A SoC With FPGA Landmark Acquisition System for Binocular Visual SLAM 12

A SoC With FPGA Landmark Acquisition System for Binocular Vi...

引用

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

作者： Schulz, Victor Hugo Bombardelli, Felipe Gustavo Todt, Eduardo Univ Fed Parana Dept Informat VRI Curitiba Parana Brazil

ISBN: (纸本)9781467371292

The present paper analyzes the use of an embedded system-on-a-chip (SoC) platform, integrating a multicore ARM processor with FPGA fabric in a single chip, as a stereo vision pre-processing module, used to retrieve depth information from the features to compose 3D landmark points. The Harris and Stephens corner detector is applied to an image pair acquired from a stereo camera setup using a hardware co-processor synthesized in the FPGA to speed up feature extraction and relieve this highly parallelizable process from the main embedded processor. The other tasks that compose the stereo vision processing module, such as image correction from camera calibration, finding a unique descriptor for the detected features, determining the correspondence between POIs in the stereo pair and retrieving depth are solved in software running on the main processor. The proposed hardware implementation enabled the corner extraction task to be performed in about half the time taken solely by the embedded ARM processor, speeding up the vision pre-processing task as a whole.

关键词： coprocessors computer vision feature extraction field programmable gate arrays robotics and automation simultaneous localization and mapping stereo vision

来源：评论

学校读者我要写书评

暂无评论

Streaming and Exploration of Dynamically Changing Dense 3D Reconstructions in Immersive Virtual Reality

Streaming and Exploration of Dynamically Changing Dense 3D R...

引用

IEEE International symposium on Mixed and Augmented Reality Workshops (ISMARW)

作者： Annette Mossel Manuel Kroeter Interactive Media Systems Group Vienna University of Technology Austria Technische Universitat Wien Wien Wien AT

ISBN: (纸本)9781509037414

We introduce a novel framework that enables large-scale dense 3D scene reconstruction, data streaming over the network and immersive exploration of the reconstructed environment using virtual reality. The system is operated by two remote entities, where one entity - for instance an autonomous aerial vehicle - captures and reconstructs the environment as well as transmits the data to another entity - such as human observer - that can immersivly explore the 3D scene, decoupled from the view of the capturing entity. The performance evaluation revealed the framework's capabilities to perform RGB-D data capturing, dense 3D reconstruction, streaming and dynamic scene updating in real time for indoor environments up to a size of 100m2, using either a state-of-the-art mobile computer or a workstation. Thereby, our work provides a foundation for enabling immersive exploration of remotely captured and incrementally reconstructed dense 3D scenes, which has not shown before and opens up new research aspects in future.

关键词： Three-dimensional displays image reconstruction Solid modeling Surface reconstruction graphics processing units Cameras Streaming media

来源：评论

学校读者我要写书评

暂无评论

Streaming Ultra High Resolution images to Large Tiled Display at Nearly Interactive Frame Rate with vl3 5

Streaming Ultra High Resolution Images to Large Tiled Displa...

引用

IEEE 5th symposium on Large Data Analysis and Visualization (LDAV)

作者： Jiang, Jie Hereld, Mark Insley, Joseph Papka, Michael E. Rizzi, Silvio Uram, Thomas Vishwanath, Venkatram Univ Illinois Chicago IL 60607 USA Argonne Natl Lab Argonne IL 60439 USA Northern Illinois Univ De Kalb IL USA

ISBN: (纸本)9781467385176

Visualization of large-scale simulations running on supercomputers requires ultra-high resolution images to capture important features in the data. In this work, we present a system for streaming ultra-high resolution images from a visualization cluster to a remote tiled display at nearly interactive frame rates. vl3, a modular framework for large scale data visualization and analysis, provides the backbone of our implementation. With this system we are able to stream over the network volume renderings of a 2048(3) voxel dataset at a resolution of 6144x3072 pixels with a frame rate of approximately 3.3 frames per second.

关键词： I.3.1 [Computing Methodologies]: computer graphics ??? Hardware Architecture: Parallel processing I.3.2 [Computing Methodologies]: computer graphics ??? graphics Systems: Distributed/network graphic Frame Rate image resolution Streaming Resolving power supercomputer Visualization Interactive images pixel Frame High definition

来源：评论

学校读者我要写书评

暂无评论

Dynamic SfM: Detecting Scene Changes from image Pairs

引用

computer graphics FORUM 2015年第5期34卷 177-189页

作者： Wang, Tuanfeng Y. Kohli, Pushmeet Mitra, Niloy J. UCL London WC1E 6BT England Microsoft Res Cambridge Cambridge England

Detecting changes in scenes is important in many scene understanding tasks. In this paper, we pursue this goal simply from a pair of image recordings. Specifically, our goal is to infer what the objects are, how they are structured, and how they moved between the images. The problem is challenging as large changes make point-level correspondence establishment difficult, which in turn breaks the assumptions of standard Structure-from-Motion (SfM). We propose a novel algorithm for dynamic SfM wherein we first generate a pool of potential corresponding points by hypothesizing over possible movements, and then use a continuous optimization formulation to obtain a low complexity solution that best explains the scene recordings, i.e., the input image pairs. We test the algorithm on a variety of examples to recover the multiple object structures and their changes.

关键词：

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：