Analysis of biomedical images requires attention to image features that represent a small fraction of the total image size. A rapid method for eliminating unnecessary detail, analogous to pre-attentive processing in b...
详细信息
ISBN:
(纸本)9781424432950
Analysis of biomedical images requires attention to image features that represent a small fraction of the total image size. A rapid method for eliminating unnecessary detail, analogous to pre-attentive processing in biological vision, allows computational resources to be applied where most needed for higher-level analysis. In this report we describe a method for bottom up merging of pixels into larger units based on flexible saliency criteria using a method similar to structured adaptive grid methods used for solving differential equations on physical domains. While creating a multiscale quadtree representation of the image, a saliency test is applied to prune the tree to eliminate unneeded details, resulting in an image with adaptive resolution. This method may be used as a first step for image segmentation and analysis and is inherently parallel, enabling implementation on programmable hardware or distributed memory clusters.
Query optimization is the most critical phase in query processing. In this paper, we try to describe synthetically the evolution of query optimization methods from uniprocessor relational database systems to data Grid...
详细信息
Three-dimensional (3D) geophysical imaging is now receiving considerable attention for electrical conductivity mapping of potential offshore oil and gas reservoirs. The imaging technology employs controlled source ele...
详细信息
Three-dimensional (3D) geophysical imaging is now receiving considerable attention for electrical conductivity mapping of potential offshore oil and gas reservoirs. The imaging technology employs controlled source electromagnetic (CSEM) and magnetotelluric (MT) fields and treats geological media exhibiting transverse anisotropy. Moreover when combined with established seismic methods, direct imaging of reservoir fluids is possible. Because of the size of the 3D conductivity imaging problem, strategies are required exploiting computational parallelism and optimal meshing. The algorithm thus developed has been shown to scale to tens of thousands of processors. In one imaging experiment, 32,768 tasks/processors on the IBM Watson Research Blue Gene/L supercomputer were successfully utilized. Over a 24 hour period we were able to image a large scale field data set that previously required over four months of processing time on distributed clusters based on Intel or AMD processors utilizing 1024 tasks on an InfiniBand fabric. Electrical conductivity imaging using massively parallel computational resources produces results that cannot be obtained otherwise and are consistent with timeframes required for practical exploration problems.
This paper presents an overview of low level parallelimageprocessing algorithms and their implementation for active vision systems. Authors have demonstrated novel low level imageprocessing algorithms for point ope...
详细信息
This paper presents an overview of low level parallelimageprocessing algorithms and their implementation for active vision systems. Authors have demonstrated novel low level imageprocessing algorithms for point operators, local operators, dithering, smoothing, edge detection, morphological operators, image segmentation and image compression. The algorithms have been prepared & described as pseudo codes. These algorithms have been simulated using parallel Computing Toolboxtrade (PCT) of MATLAB. The PCT provides parallel constructs in the MATLAB language, such as parallel for loops, distributed arrays and message passing & enables rapid prototyping of parallel code through an interactive parallel MATLAB session.
Analysis of biomedical images requires attention to image features that represent a small fraction of the total image size. A rapid method for eliminating unnecessary detail, analogous to pre-attentive processing in b...
详细信息
Analysis of biomedical images requires attention to image features that represent a small fraction of the total image size. A rapid method for eliminating unnecessary detail, analogous to pre-attentive processing in biological vision, allows computational resources to be applied where most needed for higher-level analysis. In this report we describe a method for bottom up merging of pixels into larger units based on flexible saliency criteria using a method similar to structured adaptive grid methods used for solving differential equations on physical domains. While creating a multiscale quadtree representation of the image, a saliency test is applied to prune the tree to eliminate unneeded details, resulting in an image with adaptive resolution. This method may be used as a first step for image segmentation and analysis and is inherently parallel, enabling implementation on programmable hardware or distributed memory clusters.
In this paper, we proposed 2D FFT for 8times8 matrix without transpose of data by using multiple topology on 4times4 Torus. The proposed 2D FFT used parallel operation on 1D FFT and applied an effective calculation by...
详细信息
ISBN:
(纸本)9781424445219
In this paper, we proposed 2D FFT for 8times8 matrix without transpose of data by using multiple topology on 4times4 Torus. The proposed 2D FFT used parallel operation on 1D FFT and applied an effective calculation by executing a pipeline operation. We implement the proposed architecture on Xilinx Virtex-iv device and a detailed evaluation has been reported based on maximum system frequency, chip area and image size. The implementation results show that the core speed of the proposed FFT architecture is around 157.3 MHz and it occupies 11733 slices. The average SQNR for various images is 61.9 dB. To compare the proposed 2D FFT with other methods, we can see that frame per second is improved 8 times.
The LNCS journal "Transactions on Computational Science" reflects recent developments in the field of Computational Science, conceiving the field not as a mere ancillary science but rather as an innovative a...
详细信息
ISBN:
(数字)9783642106491
ISBN:
(纸本)9783642106484
The LNCS journal "Transactions on Computational Science" reflects recent developments in the field of Computational Science, conceiving the field not as a mere ancillary science but rather as an innovative approach supporting many other scientific disciplines. The journal focuses on original high-quality research in the realm of computational science in parallel and distributed environments, encompassing the facilitating theoretical foundations and the applications of large-scale computations and massive data processing. It addresses researchers and practitioners in areas ranging from aerospace to biochemistry, from electronics to geosciences, from mathematics to software architecture, presenting verifiable computational methods, findings and solutions and enabling industrial users to apply techniques of leading-edge, large-scale, high performance computational methods. The fifth volume of the "Transactions on Computational Science" journal, edited by Yingxu Wang and Keith C.C. Chan, is devoted to the subject of cognitive knowledge representation. This field of study focuses on the internal knowledge representation mechanisms of the brain and how these can be applied to computer science and engineering. The issue includes the latest research results in internal knowledge representation at the logical, functional, physiological, and biological levels and describes their impacts on computing, artificial intelligence, and computational intelligence.
The paper is directed to the advanced Quasi Monte Carlo methods for realistic image synthesis. We propose and consider a new Quasi Monte Carlo solution of the rendering equation by uniform quadrangle separation of int...
The paper is directed to the advanced Quasi Monte Carlo methods for realistic image synthesis. We propose and consider a new Quasi Monte Carlo solution of the rendering equation by uniform quadrangle separation of integration domain. The hemispherical integration domain is uniformly separated into 12 equal size and symmetric sub‐domains. Each sub‐domain represents a solid angle, subtended by spherical quadrangle, very similar by form to plane unit square. Any spherical quadrangle has fixed vertices and computable parameters. A bijection of unit square into spherical quadrangle is find and the symmetric sampling scheme is applied to generate the sampling points uniformly distributed over hemispherical integration domain. Then, we apply the stratified Quasi Monte Carlo integration method for solving the rendering equation. The estimate of the rate of convergence is obtained. We prove the superiority of the proposed Quasi Monte Carlo solution of the rendering equation for arbitrary dimension of the sampling points. The uniform separation leads to convergence improvement of the Plain (Crude) Quasi Monte Carlo method.
暂无评论