• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11a
• /
• pp.1042-1044
• /
• 2005

고화질 입체 영상의 효과적인 재생을 위해 PC 클러스터를 활용한 여러 형태의 병렬화 기법이 제안되었지만, 영상을 구성하는 객체의 분포가 균일하지 않은 경우 충분한 성능을 발휘하지 못하였다. 본 연구에서는 Maya 렌더러를 채택한 PC 클러스터 기반의 병렬 렌더링 시스템을 구축하고, 병렬화 성능을 높이기 위한 효과적인 부하 균형 기법을 개발하였다. 특히 애니메이션을 구성하는 연속 프레임 작업에서 프레임 간의 연관성(coherence)이 높다는 사실에 근거하여, 임의 프레임의 각 분할 영역에 소요된 계산량을 바탕으로 다음 프레임의 부하 분포를 예측하고 이에 맞게 각 프로세서의 작업 영역을 재조정하는 기법을 제안하였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.3
• /
• pp.369-374
• /
• 2004

A cluster merging algorithm that merges convex clusters resulted by the Fuzzy Convex Clustering(FCC) method into non-convex clusters was proposed. This was achieved by proposing a fast and reliable distance measure between two convex clusters using Support Vector Machines(SVM) to improve accuracy and speed over other existing conventional methods. In doing so, it was possible to reduce cluster number without losing its representation of the data. In this paper, results for several data sets are given to show the validity of our distance measure and algorithm.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.12
• /
• pp.1224-1228
• /
• 2010

The existing cluster-based routing protocols have some problems. Firstly, because of selecting cluster head at random, they occur a node concentration problem. Secondly, they have a low reliability for data communication due to the less consideration of node communication range. Finally, data communication overhead is greatly increased because of sending all sensor node information to sink node for constructing clusters. To solve these problems, we in this paper, propose a cluster-based routing protocol using message reception success rate. Firstly, to solve the node concentration problem, we design a cluster head selection algorithm based on node connectivity and devise cluster spliting/merging algorithms. Secondly, to guarantee data communication reliability, we use message reception success rate. Finally, to reduce data communication overhead, we use only neighbor nodes information at both cluster construction and cluster head selection.

• Journal of the Korea Computer Graphics Society
• /
• v.17 no.4
• /
• pp.11-21
• /
• 2011

This paper proposes a method for decomposing a Hangul glyph of outline fonts into its initial, medial and final components using statistical-structural information. In a font family, the positions of components are statistically consistent and the stroke relationships of a Hangul character reflect its structure. First, we create the component histograms that accumulate the shapes and positions of the same components. Second, we make pixel clusters from character image based on pixel direction probabilities and extract the candidate strokes using position, direction, size of clusters and adjacencies between clusters. Finally, we find the best structural match between candidate strokes and predefined character model by relaxation labeling. The proposed method in this paper can be used for a study on formative characteristics of Hangul font, and for a font classification/retrieval system.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.7
• /
• pp.421-428
• /
• 2006

In this paper, the way of remote installation and back-up of 3-tier structure is introduced for efficient utilizing the cluster system resources distributed at several places. Recently, cluster system is constructed as the system of over hundreds nodes under complex network system mixed with public networks and private networks. Therefore, the as installation method suitable for the large scale cluster system and the remote recovery of failure nodes are important. However the previous researches which are based on 2-tier architecture may not provide the efficient cluster installation and image back-up method when the network of cluster system is composed of several private networks and public networks. In this paper, RISE (Remote Installation Service and Environment) based on the 3-tier architecture is proposed to solve this problem. In our approach, the managing node's role is divided into the global master node (GRISE) and the local master node (LRISE) to provide the efficient initial system deployment and remote failure recovery of distributed cluster system under the various network systems. Also, LRISE's availability is ensured under the complex network environments by adopting the auto-synchronization mechanism between GRISE and LRISE. In this work, a 64-node cluster system with gigabit network system is utilized for the experiment. From the experimental result, the system image with 1.86GB data can be obtained in 5 minutes and 53 seconds and the image-based installation of 64-node system can be carried out in 17 minutes and 53 seconds.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.1 no.3
• /
• pp.69-76
• /
• 2008

This paper concerns Fuzzy Radial Basis Function Neural Network (FRBFNN) and automatic rule generation of extraction of the FRBFNN by means of mountain clustering. In the proposed network, the membership functions of the premise part of fuzzy rules do not assume any explicit functional forms such as Gaussian, ellipsoidal, triangular, etc., so its resulting fitness values (degree of membership) directly rely on the computation of the relevant distance between data points. Also, we consider high-order polynomial as the consequent part of fuzzy rules which represent input-output characteristic of sup-space. The number of clusters and the centers of clusters are automatically generated by using mountain clustering method based on the density of data. The centers of cluster which are obtained by using mountain clustering are used to determine a degree of membership and weighted least square estimator (WLSE) is adopted to estimate the coefficients of the consequent polynomial of fuzzy rules. The effectiveness of the proposed model have been investigated and analyzed in detail for the representative nonlinear function.

• Journal of KIISE:Databases
• /
• v.29 no.2
• /
• pp.128-137
• /
• 2002

In this paper, we propose the LPC+-file for efficient indexing of high-dimensional image data. With the proliferation of multimedia data, there Is an increasing need to support the indexing and retrieval of high-dimensional image data. Recently, the LPC-file (5) that based on vector approximation has been developed for indexing high-dimensional data. The LPC-file gives good performance especially when the dataset is uniformly distributed. However, compared with for the uniformly distributed dataset, its performance degrades when the dataset is clustered. We improve the performance of the LPC-file for the strongly clustered image dataset. The basic idea is to adaptively partition the data space to find subspaces with high-density clusters and to assign more bits to them than others to increase the discriminatory power of the approximation of vectors. The total number of bits used to represent vector approximations is rather less than that of the LPC-file since the partitioned cells in the LPC+-file share the bits. An empirical evaluation shows that the LPC+-file results in significant performance improvements for real image data sets which are strongly clustered.

• The KIPS Transactions:PartC
• /
• v.17C no.5
• /
• pp.427-432
• /
• 2010

It is important to efficiently elect the cluster header in Hierarchical Sensor Network, because it largely effects on the life duration of the network. Therefore, a recent research is going forward a research activity with regard to life time extension of the whole network for efficient cluster header election. In this paper, we propose the new Cluster Header Election Scheme in which the cluster is divided into Group considering Distance from a Sink, and a cluster header will be elected by node density of the Group. Also, we evaluate the performance of this scheme, and show that this proposed scheme improves network lifetime in Zigbee environment.

• Journal of Korea Multimedia Society
• /
• v.7 no.2
• /
• pp.264-271
• /
• 2004

A load-sharing algorithm must deal with load imbalance caused by characteristics of a network and heterogeneity of nodes in Internet-based clustering systems. This paper has proposed the Efficient Load-Sharing algorithm. Efficient-Load-Sharing algorithm creates a scheduler based on the WF(Weighted Factoring) algorithm and then allocates tasks by an adaptive granularity strategy and the refined fixed granularity algorithm for better performance. In this paper, adaptive granularity strategy is that master node allocates tasks of relatively slower node to faster node and refined fixed granularity algorithm is to overlap between the time spent by slave nodes on computation and the time spent for network communication. For the simulation, the matrix multiplication using PVM is performed on the heterogeneous clustering environment which consists of two different networks. Compared to other algorithms such as Send, GSS and Weighted Factoring, the proposed algorithm results in an improvement of performance by 75%, 79% and 17%, respectively.

• The KIPS Transactions:PartA
• /
• v.14A no.7
• /
• pp.443-450
• /
• 2007

Performance scalability and storage scalability become important in a large scale cluster of wireless internet proxy cache servers. Performance scalability means that the whole performance of the cluster increases linearly according as servers are added. Storage scalability means that the total size of cache storage in the cluster is constant, regardless of the number of cache servers used, if the whole cache data are partitioned and each partition is stored in each server, respectively. The Round-Robin based load balancing method generally used in a large scale server cluster shows the performance scalability but no storage scalability because all the requested URL data need to be stored in each server. The hashing based load balancing method shows storage scalability because all the requested URL data are partitioned and each partition is stored in each server, respectively. but, it shows no performance scalability in case of uneven pattern of client requests or Hot-Spot. In this paper, we propose a novel dynamic hashing method with performance and storage scalability. In a time interval, the proposed scheme keeps to find some of requested URLs allocated to overloaded servers and dynamically reallocate them to other less-loaded servers. We performed experiments using 16 PCs and experimental results show that the proposed method has the performance and storage scalability as different from the existing hashing method.