• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2000.05a
• /
• pp.1-4
• /
• 2000

본 논문에서는 패턴 데이터(pattern data) 의 분할(partitioning)위하여, 계산량의 단축할 수 있는 효과적인 퍼지 클러스터링 알고리즘(fuzzy clustering algorithm)을 제시한다. 본 논문에 제시된 알고리즘은 두 단계로 수행된다. 첫번째 단계는, 개선된 FCM(Fuzzy C-means)방법에 의해 입력 패턴틀에 대해, 단지 두 번의 반복 수행과정만을 거쳐, 충분히 많은 개수의 초기 클러스터 중 심(center)를 결정한다. 다음 단계에서는, 본 논문에 제시될 클러스터 합치기 알고리즘(cluster merging algorithm)을 통해 각 클러스터의 부피(volume)에 따라 클러스터들을 합치는 과정(merging process)을 하게 된다. 결과적으로 일정한 제한된 개수의 무정형(amorphous)의 클러스터틀의 효과적으로 표현될 수 있다. 본 논문의 마지막에 제시될 실험 결과들은 제시된 방법의 유용성을 보여줄 것이다.

• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.2
• /
• pp.75.1-75.1
• /
• 2019

We introduce a novel method for reconstructing the projected dark matter mass maps of merging galaxy clusters by applying the convolutional neural network (CNN) to their weak lensing maps. We generate synthesized grayscale images from given weak lensing maps that preserve their averaged galaxy ellipticity. We then apply them to multi-layered CNN with architectures of alternating convolution and trans-convolution filters to predict the mass maps. We train our architecture with 1,000 Subaru/Suprime-Cam mock weak lensing maps, and our method have better mass map prediction than the Kaiser-Squires method with the following three aspects: (1) better pixel-to-pixel correlation, (2) more accurate finding of density peak position, and (3) free from mass-sheet degeneracy. We also apply our method to the HST weak lensing map of the El Gordo cluster and compare our result to the previous studies.

• The KIPS Transactions:PartD
• /
• v.13D no.7 s.110
• /
• pp.857-866
• /
• 2006

Performance of an image retrieval system is usually very low because of the semantic gap between the low level feature and the high level concept in a query image. Semantically relevant images may exhibit very different visual characteristics, and may be scattered in several clusters. In this paper, we propose a content based image rertrieval approach which combines region based image retrieval and a new relevance feedback method using adaptive clustering together. Our main goal is finding semantically related clusters to narrow down the semantic gap. Our method consists of region based clustering processes and cluster-merging process. All segmented regions of relevant images are organized into semantically related hierarchical clusters, and clusters are merged by finding the number of the latent clusters. This method, in the cluster-merging process, applies r: using v principal components instead of classical Hotelling's $T_v^2$ [1] to find the unknown number of clusters and resolve the singularity problem in high dimensions and demonstrate that there is little difference between the performance of $T^2$ and that of $T_v^2$. Experiments have demonstrated that the proposed approach is effective in improving the performance of an image retrieval system.

• Journal of The Korean Astronomical Society
• /
• v.37 no.5
• /
• pp.295-297
• /
• 2004

Observations with EUVE, ROSAT, and BeepoSAX have shown that some clusters of galaxies produce intense EUV emission. These findings have produced considerable interest; over 100 papers have been published on this topic in the refereed literature. A notable suggestion as to the source of this radiation is that it is a 'warm' (106 K) intracluster medium which, if present, would constitute the major baryonic component of the universe. A more recent variation of this theme is that this material is 'warm-hot' intergalactic material condensing onto clusters. Alternatively, inverse Compton scattering of low energy cosmic rays against cosmic microwave background photons has been proposed as the source of this emission. Various origins of these particles have been posited, including an old (${\~}$Giga year) population of cluster cosmic rays; particles associated with relativistic jets in the cluster; and cascading particles produced by shocks from sub-cluster merging. The observational situation has been quite uncertain with many reports of detections which have been subsequently contradicted by analyses carried out by other groups. Evidence supporting a thermal and a non-thermal origin has been reported. The existing EUV, FUV, and optical data will be briefly reviewed and clarified. Direct observational evidence from a number of different satellites now rules out a thermal origin for this radiation. A new examination of subtle details of the EUV data suggests a new source mechanism: inverse Compton scattered emission from secondary electrons in the cluster. This suggestion will be discussed in the context of the data.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.6
• /
• pp.523-528
• /
• 2015

In this paper, we propose a new multiresolutional and dynamic approach of the EM algorithm. EM is a very popular and powerful clustering algorithm. EM, however, has problems that indexes multiresolution data and requires a priori information on a proper number of clusters in many applications, To solve such problems, the proposed EM algorithm can impose a multiresolution kd-tree structure in the E-step and allocates a cluster based on sequential data. To validate clusters, we use a merge criteria for cluster merging. We demonstrate the proposed EM algorithm outperforms for texture image segmentation.

• 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.

• The Bulletin of The Korean Astronomical Society
• /
• v.42 no.1
• /
• pp.49.2-49.2
• /
• 2017

Star forming dwarf galaxies in various environments are attractive objects for investigating the environmental effects on chemical evolution of dwarf galaxies. Using SDSS DR7 spectroscopic data and GALEX ultraviolet (UV) imaging data, we study the chemical properties of star forming dwarf galaxies in various environments of the Virgo cluster, Ursa Major group, and field. We derived gas-phase abundance, galaxy mass, and UV specific star formation rate (sSFR) of subsample, early-type (ETD) and late-type star forming dwarf (LTD) galaxies, which are divided by visually classified galaxy morphology. We found no O/H enhancement of LTDs in cluster and group environments compared to the field, implying no environmental dependence of the mass-metallicity relation for LTDs. LTDs in the Virgo cluster and Ursa Major group have similar sSFR at a given galaxy mass, but they exhibit systematically lower sSFR than those in isolated field environment. We suggest that LTDs in the Virgo cluster are an infalling population that was recently accreted from the outside of the cluster. We found that ETDs in the Virgo cluster and Ursa Major group exhibit enhanced O/H compared to those in the field. However, no distinct difference of N/O of galaxies between different environments. The chemically evolved ETDs in the Virgo cluster and Ursa Major group also show similar mass-sSFR relation, but systematically lower sSFR at a fixed galaxy mass compared to the field counterparts. We suggest that ETDs in the Virgo cluster and Ursa Major group have evolved under the similar local environments. We also discuss the evolutionary path of ETDs and LTDs with respect to the environmental effects of ram pressure stripping and galaxy interaction/merging.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2002.12a
• /
• pp.267-270
• /
• 2002

본 논문에서는 Support Vector Machines (SVM) 을 이용하여, 빠르고 정확한 두 convex한 클러스터 간의 거리 측정 방법을 제시한다 제시된 방법에서는, SVM에 의해서 생성되는 최적 다차원 평면이 두 클러스터간의 최소 거리를 계산하는데 사용된다. 또한, 본 논문에서는 이러한 두 클러스터 간의 최적의 거리를 사용하여, Fuzzy Convex Clustering (FCC) 방법 (1) 에 의해서 생성되는 Convex 클러스터들을 묶어주는 효과적인 클러스터 결합 알고리즘을 제시하였다. 그러므로, 데이터의 부적절한 표현을 유발하지 않고도 클러스터들의 개수를 좀 더 줄일 수 있었다. 제시한 방법의 타당성을 위하여 여러 실험 결과를 제시하였다

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.10a
• /
• pp.468-471
• /
• 2003

In this paper, experimental tests are performed to evaluate the efficiency of spatial clustering method using regular grid that is proposed in our recent research. In details, we estimate the execution time for finding clusters varying spatial objects on sample data sets with various distributions and perform experimental tests varying threshold value on a data set. We also compare the running time of cluster generating algorithm with that of cluster merging algorithm per each test.

• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.1
• /
• pp.49.2-49.2
• /
• 2019

Clusters of galaxies shape up through a series of hierarchical mergers. It is believed that major mergers lead to cluster-wide shock waves, which are manifested as radio relics. The 1RXS J0603.0+4213 and CIZA J2242.8+5301 clusters, for instance, contain Mpc-size giant radio relics in the outskirts. Synchrotron emission from these radio relics reveals the presence of relativistic electrons and the magnetic fields of a few ${\mu}G$ strength. The presence of such magnetic fields in the ICM has been explained by the so-called small-scale turbulent dynamo. To get quantitative measures for magnetic fields in clusters of galaxies, we investigate the development of turbulence and the follow-up amplification of magnetic fields through three-dimensional numerical magnetohydrodynamical (MHD) simulations. The turbulence is induced in highly stratified cluster media, and driven sporadically by major mergers. We here present the results, aiming to answer whether the turbulence dynamo scenario can explain the observed strength and scale of magnetic fields in clusters. Also, we discuss whether the observed properties of giant radio relics can be reproduced in our simulations.