In this paper problem of graph based image segmentation is considered. In particular, attention is paid to minimal spanning tree based algorithm proposed by Felzenszwalb and Huttenlocher (FH). Although the method yiel...
详细信息
In this paper problem of graph based image segmentation is considered. In particular, attention is paid to minimal spanning tree based algorithm proposed by Felzenszwalb and Huttenlocher (FH). Although the method yields high quality results for various classes of images, its application is limited mainly to off-line processing. Its due to the very long execution time of the FH method, which in the case of high resolution images, requires processing of millions of vertices and edges contained within the resulting graph. Therefore, some improvements to the FH method are proposed in this paper. The modifications aim at the reduction of algorithm execution time and the usage of computer host memory. These goals are achieved both by reducing the size of input image graph and by applying the methods of GPU parallel computing at initial stages of the algorithm. As the reduction of graph size is obtained by processing meta-pixels representing homogenous regions, the new method is most suitable for the segmentation of images including rare, structurally complex objects distributed over nonuniform background. Results obtained by the introduced approach are compared with the results of the original FH method and other popular graph-based approaches to image segmentation. The comparison includes both the accuracy of image segmentation and the execution time. Analysis of the results clearly shows, that the proposed approach in many cases can significantly accelerate segmentation process without a noticeable loss of image segmentation quality.
This paper proposes Virtual Video Tape (WT). It is a randomly accessible motion image recorder in main memory. VVT is realized with only software, not hardware. It is intended as a tool for real-time motion image unde...
详细信息
ISBN:
(纸本)0819437638
This paper proposes Virtual Video Tape (WT). It is a randomly accessible motion image recorder in main memory. VVT is realized with only software, not hardware. It is intended as a tool for real-time motion image understanding research. Recent remarkable progress of PC hardware enables to use gigabyte order main memory. By utilizing such large sized memory, there is a possibility to realize motion image recorder with software. Thus we propose VVT as an example of this kind image recorder. Utilizing current components, recording time fan be expected as minutes order. Since the proposed VVT is fully digital, there is no analog medium nor possibility for degradation of image quality. Since there is no deterioration of playback image and no rewinding. VVT must contribute to program development for motion image understanding. Based upon the proposed idea, the authors have implemented a prototype VVT and used the prototype to develop visual tracking, real-time face detection, and so forth. Through the implementation and experience of the usage, we have confirmed feasibility and effectiveness of the proposed idea. In this paper, the authors discuss background, required functions and structure of the recorder. Some implementation issues are also described.
In many numerical applications resulting from computational science and engineering problems, the solution of sparse linear systems is the most prohibitively compute intensive task. Consequently, the linear solvers ne...
详细信息
ISBN:
(纸本)9780769539393
In many numerical applications resulting from computational science and engineering problems, the solution of sparse linear systems is the most prohibitively compute intensive task. Consequently, the linear solvers need to be carefully chosen and efficiently implemented in order to harness the available computing resources. Krylov subspace based iterative solvers have been widely used for solving large systems of linear equations. In this paper, we focus on the design of such iterative solvers to take advantage of massive parallelism of general purpose Graphics processing Units (GPU)s. We will consider Stabilized BiConjugate Gradient (BiCGStab) and Conjugate Gradient Squared (CGS) methods for the solutions of sparse linear systems with unsymmetric coefficient matrices. We discuss data structures and efficient implementation of these solvers on the NVIDIA's CUDA platform. We evaluate scalability and performance of our implementations in the context of a financial engineering problem of solving multidimensional option pricing PDEs using sparse grid combination technique.
The proceedings contain 74 papers. The topics discussed include: web services composition for distributed data mining;how to run scientific applications over web services;resource management services for a grid analys...
详细信息
The proceedings contain 74 papers. The topics discussed include: web services composition for distributed data mining;how to run scientific applications over web services;resource management services for a grid analysis environment;a comparison of two methods for building astronomical image mosaics on a grid;gene sequence alignment on a public computing platform;parallel module network learning on distributed memory multiprocessors;a scalable parallel poisson solver in three dimensions with infinite-domain boundary conditions;factoring solution sets of polynomial systems in parallel;a programmable array processor architecture for flexible approximate string matching algorithms;and speculative parallel threading architecture and compilation.
Remote Sensing (RS) data processing is characterized by massive remote sensing images and increasing amount of algorithms of higher complexity. parallel programming for data-intensive applications like massive remote ...
详细信息
ISBN:
(纸本)9781467324229
Remote Sensing (RS) data processing is characterized by massive remote sensing images and increasing amount of algorithms of higher complexity. parallel programming for data-intensive applications like massive remote sensing imageprocessing on parallel systems is bound to be especially trivial and challenging. We propose a C++ template mechanism enabled generic parallel programming skeleton for these remote sensing applications in high performance clusters. It provides both programming templates for distributed RS data and generic parallel skeletons for RS algorithms. Through one-side communication primitives provided by MPI, the distributed RS data template could provide a global view of the big RS data whose sliced data blocks are scattered among the distributed memory of cluster nodes. Moreover, by data serialization and RMA (Remote Memory Access), the data templates could also offer a simple and effective way to distribute and communicate massive remote sensing data with complex data structures. Furthermore, the generic parallel skeletons implement the recurring patterns of computation, performance optimization and pass the user-defined sequential functions as parameters of templates for type genericity. With the implemented skeletons, Developers without extensive parallel computing technologies can implement efficient parallel remote sensing programs without concerning for parallel computing details. Through experiments on remote sensing applications, we confirmed that our templates were productive and efficient.
For the solutions of large and sparse linear systems of equations with unsymmetric coefficient matrices, we propose an improved version of the Conjugate Gradient Squared method (ICGS) method. The algorithm is derived ...
详细信息
ISBN:
(纸本)0769512607
For the solutions of large and sparse linear systems of equations with unsymmetric coefficient matrices, we propose an improved version of the Conjugate Gradient Squared method (ICGS) method. The algorithm is derived such that all inner products, matrix-vector multiplications and vector updates of a single iteration step are independent and communication time required for inner product can be overlapped efficiently with computation time of vector updates. Therefore, the cost of global communication on paralleldistributed memory computers can be significantly reduced. The resulting ICGS algorithm maintains the favorable properties of the algorithm while not increasing computational costs. Data distribution suitable for both irregularly and regularly structured matrices based on the analysis of the non-zero matrix elements is also presented. Communication scheme is supported by overlapping execution of computation and communication to reduce waiting times. The efficiency of this method is demonstrated by numerical experimental results carried out on a massively paralleldistributed memory system.
The finite-difference time-domain (FDTD) method is the well-known method to solve three-dimensional (3-D) electromagnetic problems. However, it contains numerical instability because it is a complete explicit method, ...
详细信息
The finite-difference time-domain (FDTD) method is the well-known method to solve three-dimensional (3-D) electromagnetic problems. However, it contains numerical instability because it is a complete explicit method, and huge memory is required in the case of 3-D analysis. In this paper, the implicit FDTD method called the alternating implicit block overlapped (AIBO)-FDTD method is applied to a 3-D electromagnetic analysis. This method is numerically stable since the implicit scheme is included, and it is suitable for parallelprocessing since the computational domain can be effectively partitioned. This property is especially compatible with implementations using distributed-memory multiprocessor systems such as PC clusters, which provide a huge memory space that accommodate a large problem. Our results show that parallelprocessing is an effective technique for improving the performance of the 3-D AIBO-FDTD method, while this method is slower than FDTD when it is run on one processor.
Nowadays, image compression is one of the essential tool used in the digital imageprocessing. Data compression is used to minimize the quantity as well as undesirable information while maintaining useful information ...
详细信息
ISBN:
(纸本)9781509036691
Nowadays, image compression is one of the essential tool used in the digital imageprocessing. Data compression is used to minimize the quantity as well as undesirable information while maintaining useful information within the image. Transform based compression can be broadly used by image compression. However, transform based techniques expose obstructing artifacts within the resultant image. This paper recommends the new technique as a way to minimize obstructing artifacts within compressed image. The latest technique combined SVD-WDR compression along with Gradient-based optimization method for removing blocking and ringing artifacts from the compressed images.
This paper presents a software framework providing a platform for parallel and distributedprocessing of video data oil a cluster of SMP computers. Existing video-processing algorithins can be easily integrated into t...
详细信息
ISBN:
(纸本)9780819466099
This paper presents a software framework providing a platform for parallel and distributedprocessing of video data oil a cluster of SMP computers. Existing video-processing algorithins can be easily integrated into the framework by considering them as atomic processing tiles (PTs). PTs can be connected to form processing graphs that model the data flow of a specific application. This graph also defines the data dependencies that determine using which tasks can be computed in parallel. Scheduling of the tasks in this graph is carried out automatically using a pool-of-tasks scheme. The data format that can be processed by the framework is not restricted to image data, such that also intermediate data, like detected feature points or object positions, can be transferred between PTs. Furthermore, the processing can optionally be carried out efficiently on special-purpose processors with separate memory, since the framework minimizes the transfer of data. Finally, we describe an example application for a multi-camera view-interpolation system that we successfully implemented on the proposed framework.
The processing of three-dimensional (3-D) objects from 3-D digital image data is an important task in the imageprocessing and the computer vision fields. The distance transform (DT) is extensively applied in the imag...
详细信息
ISBN:
(纸本)0780378407
The processing of three-dimensional (3-D) objects from 3-D digital image data is an important task in the imageprocessing and the computer vision fields. The distance transform (DT) is extensively applied in the imageprocessing and computer vision areas as a key operation. In a two or three-dimensional image array, the computation of distance transform (DT) is an important task, With the increasing application of 3D voxel images, it is useful to consider the distance transform of a 3D digital image array. In order to provide the efficient transform computations, parallelism is employed. We develop parallel algorithms for the three-dimensional Euclidean distance transform (3D-EDT) on the SIMD hypercube computer. The time complexity of our parallel algorithm is O{log(2) N} for an N x N x N image array using N-3 processors. A generalized parallel algorithm for the 3D-EDT is also proposed and it runs O((N/p)(3) log(N) + (N/p)(2) log(2) p) time for an N x N x N binary image array on the SIMD hypercube computer using p(3) PE's, where 1 less than or equal to p less than or equal to N.
暂无评论