Embedded multimedia systems are expected to fully embrace the future many-core wave. As a consequence parallel programming is being revamped as the only way to exploit the power of coming chips. While waiting for them...
详细信息
The aim of the paper is to validate a software architecture that allows an imageprocessing researcher to develop parallel applications. The challenge was to develop algorithms that perform real-time low level operati...
详细信息
ISBN:
(纸本)9789604741762
The aim of the paper is to validate a software architecture that allows an imageprocessing researcher to develop parallel applications. The challenge was to develop algorithms that perform real-time low level operations on digital images able to be executed on a cluster of desktop PCs. The experiments show how to use parallelizable patterns and how to optimize the load balancing between the workstations.
This paper proposes a triangular solve algorithm with variable block size for graphics processing unit (GPU). By using diagonal blocks inversion with recursion, this algorithm works with tunable block size to achieve ...
详细信息
ISBN:
(纸本)9780769541570
This paper proposes a triangular solve algorithm with variable block size for graphics processing unit (GPU). By using diagonal blocks inversion with recursion, this algorithm works with tunable block size to achieve the best performance. Various methods are shown on how to make use of existing profiling tools to successfully measure and analyze performance of this *** use some of the most popular CPU and GPU profiling tools for their advantages and overcome their disadvantages with several new techniques to analyze the performance and relationship of different components of applications. With the presented methodologies, insight information is produced which helps to understand and tune the proposed algorithm and considerably improve the performance of the solver itself as well as the application using it.
Room reverberation leads to reduced intelligibility of audio signals. Enhancement is thus crucial for high-quality audio and scene analysis applications. This paper proposes to directly and optimally estimate the sour...
详细信息
Room reverberation leads to reduced intelligibility of audio signals. Enhancement is thus crucial for high-quality audio and scene analysis applications. This paper proposes to directly and optimally estimate the source signal and acoustic channel from the distorted observations. The remaining model parameters are sampled from a particle filter, facilitating real-time dereverberation. The approach was previously successfully applied to single- and multi-sensor blind dereverberation. Enhancement can be improved upon by accurately modelling the speech production system. This paper therefore extends the blind dereverberation approach to incorporate a novel source model based on parallel formant synthesis and compares the approach to one using a time-varying AR model, with parameters varying according to a random walk. Experimental data shows that dereverberation using the proposed model is improved for vowels, stop consonants, and fricatives.
Color blindness is a serious perception problem. Suffering individuals cannot understand messages, which are carrying by many images, especially those, distributed by WWW. In this paper we are proposing three image pr...
详细信息
As an extension of traditional data mining, distributed data mining (DDM) is a well-known issue that has been extensively studied to extract some hidden, useful and interesting information from data by using various k...
详细信息
ISBN:
(纸本)9780769539515
As an extension of traditional data mining, distributed data mining (DDM) is a well-known issue that has been extensively studied to extract some hidden, useful and interesting information from data by using various kinds of resources distributed in different geography positions. In the literature, there are many methods and technologies proposed to implement large-scale distributed data mining, but owing to lack of autonomy, intelligentization and the limitation of expansibility, most of the works fail to reach a practical application in industry and commerce fields. With the rapid development of distributed sharing infrastructures and the emergence and mature of grid technology, providing knowledge services in grid environment is becoming the trend of knowledge discovery, and it will certainly make the technology of data processing into a broader way. In this paper, we elaborate on a geographical knowledge services grid test-bed, which is developed to support distributed & parallel geological problems solving, as an example, a distributed spatial outliers mining algorithm based on Delaunay triangulation(D-TIN) is introduced, and the perform scheduling of Delaunay triangulation construction and spatial outliers mining is presented too.
We show a novel execution method of queries over structural data. We present the idea in detail on SBQL (a.k.a. AOQL)-a powerful language with clean semantics. SBQL stands for the Stack-Based Query Language. The stack...
详细信息
ISBN:
(数字)9783642176258
ISBN:
(纸本)9783642176241
We show a novel execution method of queries over structural data. We present the idea in detail on SBQL (a.k.a. AOQL)-a powerful language with clean semantics. SBQL stands for the Stack-Based Query Language. The stack used in its name and semantics is a heavy and centralised structure which makes parallel and stream processing unfeasible. We propose to process stack-based queries without a stack. The stack-less execution of stack-based queries is done by a network of components connected by streams. Source streams of the network are populated with necessary fragments of objects retrieved from the database. Output streams produce results of the query. This paper focuses on methods of construction, optimisation and effective execution of such a networks. Although this research was inspired by SBQL/AOQL, eventually we achieved a universal execution method for other object-oriented, object-relational and semi-structured language. Especially promising are the optimisation and parallelisation possibilities offered by the presented execution model.
Computing image local statistics is required in many imageprocessing applications such as local adaptive image restoration, enhancement, segmentation, target location and tracking, to name a few. These computations m...
详细信息
ISBN:
(纸本)9780819481979
Computing image local statistics is required in many imageprocessing applications such as local adaptive image restoration, enhancement, segmentation, target location and tracking, to name a few. These computations must be carried out in sliding window of a certain shape and weights. Generally, it is a time consuming operation with per-pixel computational complexity of the order of the window size, which hampers real-time applications. For acceleration of computations, recursive computational algorithms are used. However, such algorithms are available only for windows of certain specific forms, such as rectangle and octagon, with uniform weights. We present a general framework of fast parallel and recursive computation of image local statistics in sliding window of almost arbitrary shape and weights with "per-pixel" computational complexity that is substantially of lower order than the window size. As an illustration of this framework, we describe methods for computing image local moments such as local mean and variance, image local histograms and local order statistics (in particular, minimum, maximum, median), image local ranks, image local DFT, DCT, DcST spectra in polygon-shaped windows as well as in windows with non-uniform weights, such as Sine lobe, Hann, Hamming and Blackman windows.
As the image size increases, it is difficult to perform the large size imageprocessing on a single computer in a limited time. One of the most efficient solving methods is parallel computing. Most numerical approache...
详细信息
distributed deductive databases utilize multiple processing nodes to complete their tasks more efficiently. Efficiencies can also be gained by optimizing the way that the predicates are fired. One way of doing this is...
详细信息
ISBN:
(纸本)9781601321480
distributed deductive databases utilize multiple processing nodes to complete their tasks more efficiently. Efficiencies can also be gained by optimizing the way that the predicates are fired. One way of doing this is to activate all of the predicates at the same time. A second way is to optimize the activation order of the predicates. This paper presents both of these approaches and then introduces an optimization strategy for which approach to use by examining some of the unique parameters of each database.
暂无评论