In this paper, a new method for line detection at sport images using multi-agent systems is proposed. For this purpose, we introduced some agents. Each agent can recognize special section of an image, for example one ...
详细信息
ISBN:
(纸本)9780769535913
In this paper, a new method for line detection at sport images using multi-agent systems is proposed. For this purpose, we introduced some agents. Each agent can recognize special section of an image, for example one agent for recognition of sport image's background (green pixels) and another for recognition of images line (white pixels). The results show good performance with respect to other methods, for example, Hough algorithm. This method has many advantages such as simplicity and high speed implementation, availability of line and background separately, cancelation of unnecessary sections such as viewer sections and parallel implementation ability.
This paper presents an efficient parallel implementation on Graphics processing Units (GPUs) for the Simplified PN (SPN) calculations in the 3D case. For a nuclear operator such as EDF, the time required to compute nu...
详细信息
ISBN:
(纸本)9781615673490
This paper presents an efficient parallel implementation on Graphics processing Units (GPUs) for the Simplified PN (SPN) calculations in the 3D case. For a nuclear operator such as EDF, the time required to compute nuclear reactor core simulations is rather critical. The SPN method provides a convenient trade-off between accuracy and numerical complexity and is used in several industrial simulations. The parallelization of the algorithm should allow to reduce the computation time required to solve the eigenvalue problem. To solve the problem on distributed memory machines such as PC clusters, Domain Decomposition methods have been investigated. Complementary to this approach, this work aims at using emerging massively parallel processors such as the GPUs. Based on a fine grained parallelism, this solution offers the opportunity to achieve good performances at very low cost. Indeed, GPUs provide a large computational power and some specific optimizations allow to near the hardware limits. Our GPU implementation solves 3D SPN problems 30 times faster than its sequential CPU counterpart.
Microscopic imaging is an important tool for characterizing tissue morphology and pathology. 3D reconstruction and visualization of large sample tissue structure requires registration of large sets of high-resolution ...
详细信息
Microscopic imaging is an important tool for characterizing tissue morphology and pathology. 3D reconstruction and visualization of large sample tissue structure requires registration of large sets of high-resolution images. However, the scale of this problem presents a challenge for automatic registration methods. In this paper we present a novel method for efficient automatic registration using graphics processing units (GPUs) and parallel programming. Comparing a C++ CPU implementation with Compute Unified Device Architecture (CUDA) libraries and pthreads running on GPU we achieve a speed-up factor of up to 4.11x with a single GPU and 6.68x with a GPU pair. We present execution times for a benchmark composed of two sets of large-scale images: mouse placenta (16K x16K pixels) and breast cancer tumors (23K x62K pixels). It takes more than 12 hours for the genetic case in C++ to register a typical sample composed of 500 consecutive slides, which was reduced to less than 2 hours using two GPUs, in addition to a very promising scalability for extending those gains easily on a large number of GPUs in a distributed system.
In this paper, we propose a Bayesian model and a Monte Carlo Markov chain (MCMC) algorithm for reconstructing images that consist of only few non-zero pixels. An appropriate distribution that promotes sparsity is prop...
详细信息
ISBN:
(纸本)9781424423538
In this paper, we propose a Bayesian model and a Monte Carlo Markov chain (MCMC) algorithm for reconstructing images that consist of only few non-zero pixels. An appropriate distribution that promotes sparsity is proposed as prior distribution for the pixel values. The hyperparameters involved in the modeling are also assigned prior distributions, resulting in a hierarchical model. A Gibbs sampler allows us to draw samples distributed according the full posterior of interest. These samples are then used to approximate standard maximum a posteriori (MAP) estimator. By conducting some simulations, we show that the proposed estimator clearly outperforms previous estimators proposed in the literature.
Several regular parallel trees have been proposed over the years to optimize logic depth, area, fan-out and interconnect count for logic circuits. In this paper, we propose a comparative study of different parallel pr...
详细信息
ISBN:
(纸本)9780769537825
Several regular parallel trees have been proposed over the years to optimize logic depth, area, fan-out and interconnect count for logic circuits. In this paper, we propose a comparative study of different parallel prefix trees used in the design of a new end-around carry (EAC) adder targeting FPGA technology. This new adder is based on the fast 128-bit binary floating-point EAC adder which has been implemented in the IBM POWER6 microprocessor's fused multiply-add unit. The parallel prefix tree implemented on the IBM's EAC adder is a Kogge-Stone tree which has been chosen for its high performance and its low power consumption. Our comparative study highlights the main performance differences among fourteen different architecture configurations when targeting an FPGA EAC adder design. We focus on the area requirements and the critical path delay of these designs. Our experimental results show that there is one architecture configuration with the lower area requirement and the higher performance.
This article describes a new architecture for a parallel, digital image processor which performs several imageprocessing tasks like segmentation, edge detection and noise removal. The architecture and algorithm modif...
详细信息
Tissue P systems are a class of distributed and parallel computing models inspired by intercellular communication and cooperation between neurons. An interesting variant of tissue P systems is known as tissue P system...
详细信息
ISBN:
(纸本)9781424438655
Tissue P systems are a class of distributed and parallel computing models inspired by intercellular communication and cooperation between neurons. An interesting variant of tissue P systems is known as tissue P systems with cell separation, which is endowed with the ability of generating an exponential workspace in a linear time and in this way getting the possibility to solve computationally hard problems in polynomial time. In this paper, a uniform solution to 3-coloring problem by tissue P systems with cell separation is presented, which is linear in terms of the number of vertices and the number of edges.
The process of image formation in an optical microscope or similar imaging device results in a distortion of the true object image due to diffraction effects and out-of-focus blurring. This distortion can greatly limi...
详细信息
ISBN:
(纸本)9780975840078
The process of image formation in an optical microscope or similar imaging device results in a distortion of the true object image due to diffraction effects and out-of-focus blurring. This distortion can greatly limit the resolution of the imaging device, particularly in the case of 3D microscopy where the axial resolution is impeded by the contribution of out-of-focus signal from an extended area of the object outside a recorded focal plane. Fortunately the image formation process can be modelled as a convolution of the original object with the impulse response of the device, that is, the image of a point energy source. As such, an approximation of the original object image can be derived through an inversion of this model - a deconvolution. Unfortunately the inclusion of random noise in the image recording process makes direct methods of inverting this model less than ideal as they are prone to amplification of noise. A class of popular approaches is to use iterative methods which try to account for the noise and progress towards an acceptable approximation of the real image. These methods are often computationally intensive, requiring a reapplication of the image formation model to successive estimates of the real image. In addition, biological and medical research organisations are collecting 3D image data on an increasingly large scale, and there is a demand to process this data in a timely manner. In this paper we investigate the use of graphics processing units (GPUs) to accelerate the execution of one such iterative algorithm, the Richardson-Lucy (RL) algorithm. Modern GPUs are highly parallel commodity processors containing hundreds of cores and capable of executing thousands of threads concurrently. GPUs can be utilised to accelerate a wide variety of compute intensive algorithms. As their programmability has improved over the past decade, significant effort has been invested in performing general purpose computing on GPUs (GPGPU). Until recently GPGPU al
The availability of huge-scale computing platforms comprised of tells of thousands of multicore processors motivates the need for the next generation of highly scalable sparse linear system solvers. These solvers must...
详细信息
ISBN:
(纸本)9783642038686
The availability of huge-scale computing platforms comprised of tells of thousands of multicore processors motivates the need for the next generation of highly scalable sparse linear system solvers. These solvers must optimize parallel performance, processor (serial) performance;as well as memory requirements, while being robust across broad classes of applications and systems. In this paper, we present;a new parallel solver that combines the desirable characteristics of direct methods (robustness) and effective iterative solvers (low computational cost), while alleviating their drawbacks (memory requirements, lack of robustness). Our proposed hybrid solver is based on tire general sparse solver PARDISO;and the "Spike" family of hybrid solvers. The resulting algorithm, called PSPIKE, is as robust as direct solvers, more reliable than classical preconditioned Krylov subspace methods, and much more scalable than direct sparse solvers. We support;our performance and parallel scalability claims using detailed experimental studies and comparison with direct solvers, as well as classical preconditioned Krylov methods.
暂无评论