image registration is to find a suitable space transformation, making two or more images match accurately on the same points. The main time consuming is on the matrix computation over the iterative procession. As the ...
详细信息
image registration is to find a suitable space transformation, making two or more images match accurately on the same points. The main time consuming is on the matrix computation over the iterative procession. As the image size increasing, the numerical approaches based on the pixel matrix may step up the computation complexity, and it is difficult to perform the large size registration on a single computer in a limited time. Then one of the most efficient solving methods is parallel computing. In this paper, we discuss a multithreading method to produce the image registration computing, detail its parallel implementation on a 4 nodes dual-core cluster, and apply this procession to the meteorological objective tracking. Some experimental results show that this method can produce the large size image registration and reduce the computation time.
One among the most influential and popular data mining methods is the k-Means algorithm for cluster analysis. Techniques for improving the efficiency of k-Means have been largely explored in two main directions. The a...
详细信息
One among the most influential and popular data mining methods is the k-Means algorithm for cluster analysis. Techniques for improving the efficiency of k-Means have been largely explored in two main directions. The amount of computation can be significantly reduced by adopting geometrical constraints and an efficient data structure, notably a multidimensional binary search tree (KD-Tree). These techniques allow to reduce the number of distance computations the algorithm performs at each iteration. A second direction is parallelprocessing, where data and computation loads are distributed over many processing nodes. However, little work has been done to provide a parallel formulation of the efficient sequential techniques based on KD-Trees. Such approaches are expected to have an irregular distribution of computation load and can suffer from load imbalance. This issue has so far limited the adoption of these efficient k-Means variants in parallel computing environments. In this work, we provide a parallel formulation of the KD-Tree based k-Means algorithm for distributed memory systems and address its load balancing issue. Three solutions have been developed and tested. Two approaches are based on a static partitioning of the data set and a third solution incorporates a dynamic load balancing policy.
Development and implementation of the coordination language S-NET has been reported previously. In this paper we apply the S-NET design methodology to a computer graphics problem. We demonstrate (i) how a complete sep...
详细信息
Development and implementation of the coordination language S-NET has been reported previously. In this paper we apply the S-NET design methodology to a computer graphics problem. We demonstrate (i) how a complete separation of concerns can be achieved between algorithm engineering and concurrency engineering and (ii) that the S-NET implementation is quite capable of achieving performance that matches what can be achieved using low-level tools such as MPI. We find this remarkable as under S-NET communication, concurrency and synchronization are completely separated from algorithmic code. We argue that our approach delivers a flexible component technology which liberates application developers from the logistics of task and data management while at the same time making it unnecessary for a distributed computing professional to acquire detailed knowledge of the application area.
A novel multi-view camera system based on distributed Video Coding (DVC) scheme is described in this paper. Unlike other systems that use camera arrays, the proposed system consists of multiple uncalibrated and vastly...
详细信息
A novel multi-view camera system based on distributed Video Coding (DVC) scheme is described in this paper. Unlike other systems that use camera arrays, the proposed system consists of multiple uncalibrated and vastly different still-image cameras. The relative positions and viewing angles of these uncalibrated cameras are not known a-priori. images captured by these cameras are first registered using sparse, low level image features. Inter-view correlation is then exploited iteratively using a DVC scheme. Our results demonstrated that it is feasible to generate high-quality multi-view images from different and uncalibrated cameras based on DVC methods.
Nowadays, large scale distributed systems gather thousands of nodes with hierarchical memory models. They are heterogeneous, volatile and geographically distributed. The efficient exploitation of such systems requires...
详细信息
Nowadays, large scale distributed systems gather thousands of nodes with hierarchical memory models. They are heterogeneous, volatile and geographically distributed. The efficient exploitation of such systems requires the conception and adaptation of appropriate numerical methods, the definition of new programming paradigms, new metrics for performance prediction, etc. The modern hybrid numerical methods are well adapted to this kind of systems. This is particularly because of their multi-level parallelism and fault tolerance property. However the programming of these methods for these architectures requires concurrent reuse of sequential and parallel code. But the currently existing numerical libraries aren't able to exploit the multi-level parallelism offered by theses methods. A few linear algebra numerical libraries make use of object oriented approach allowing modularity and extensibility. Nevertheless, those which offer modularity,sequential and parallel code reuse are almost non-existent. In this paper, we analyze the lacks in existing libraries and propose a design based on a component approach and the strict separation between computation operations, data management and communication control of an application. We present then an application of this design using YML scientific workflow environment (http://***/) jointly with the object oriented LAKe (Linear Algebra Kernel) library. Some numerical experiments on GRID5000 platform validate our approach and show its efficiency.
The multi-objective integer programming problems in large scale are considered time consuming. In the past, mathematical structures were used that can get benefits of high processing powers and parallelprocessing. Th...
详细信息
ISBN:
(纸本)9781424458288;9789774033964
The multi-objective integer programming problems in large scale are considered time consuming. In the past, mathematical structures were used that can get benefits of high processing powers and parallelprocessing. The Branch and Bound (B&B) algorithm is one of the most used methods to solve combinatorial optimization problems. A general approach to generate all non-dominated solutions of the multi-objective integer programming (MOIP) Problem is developed. In this paper, a supervisor-master-sub-master-worker algorithm to solve large scale integer multi-objective problems that can get the benefits of mathematical structures and high processing powers has been proposed. This approach addresses several issues related to the characteristics of the algorithm itself and the properties of parallel computing systems. From the solved benchmark example this algorithm proved to provide a considerable high performance. Results show that a consistently better efficiency can be achieved in solving integer equations, providing reduction of time.
This paper proposes a novel approach for feature extraction using bit plane slicing. This technique is preferred over some traditional methods of feature extraction using orthogonal and non orthogonal transforms that ...
详细信息
This paper proposes a novel approach for feature extraction using bit plane slicing. This technique is preferred over some traditional methods of feature extraction using orthogonal and non orthogonal transforms that uses real features. The extracted real features are computationally expensive in terms of time and storage. The proposed iris based biometric system is implemented by localizing the inner pupil and outer iris boundary using dynamic thresholding technique and Circular Hough Transform. The localized iris texture is mapped into rectangular block to extract unique descriptors. For feature extraction and iris code generation bit plane slicing approach is used. After generation of bit codes, matching is done using bitwise XOR operator. The system is tested on BATH, CASIA and IITK databases and outperforms existing approaches in terms of accuracy.
We propose a performance estimation technique for a multi-core segmented bus platform, SegBus. The technique enables us to assess the performance aspects of any specific application on a particular platform configurat...
详细信息
We propose a performance estimation technique for a multi-core segmented bus platform, SegBus. The technique enables us to assess the performance aspects of any specific application on a particular platform configuration, modeled in Unified Modeling Language (UML). We present methods to transform Packet Synchronous Data Flow (PSDF) and Platform Specific Model (PSM) models of the application into Extensible Markup Language (XML) schemes using modeling tool and how the generated XML schemes can be utilized by the emulator program to get the execution results. The technique facilitates us to estimate performance aspects of application mapped on a number of different platform configurations during the early stages of the design process.
RFID technology holds the promise of real-time identifying, locating and monitoring physical objects. Complex event processing is the key part of the RFID middleware. Most of the current methods of complex event ...
详细信息
RFID technology holds the promise of real-time identifying, locating and monitoring physical objects. Complex event processing is the key part of the RFID middleware. Most of the current methods of complex event processing can't support large-scale RFID applications since they have not considered much on performance and distributed data. In this paper, a distributed CEP architecture is proposed which spreads centralized CEP tasks load over multiple communicating stations over CORBA. Based on this architecture, a high performance plan based distributed CEP method is proposed which uses parallel operations to improve performance. The experimental study sbows that the performance of our approach is remarkable in largescale RFID applications.
暂无评论