• Journal of Information Science Theory and Practice
• /
• 제6권4호
• /
• pp.28-38
• /
• 2018

Wikipedia is composed of millions of articles, each of which explains a particular entity with various languages in the real world. Since the articles are contributed and edited by a large population of diverse experts with no specific authority, Wikipedia can be seen as a naturally occurring body of human knowledge. In this paper, we propose a method to automatically identify key entities and relations in Wikipedia articles, which can be used for automatic ontology construction. Compared to previous approaches to entity and relation extraction and/or identification from text, our goal is to capture naturally occurring entities and relations from Wikipedia while minimizing artificiality often introduced at the stages of constructing training and testing data. The titles of the articles and anchored phrases in their text are regarded as entities, and their types are automatically classified with minimal training. We attempt to automatically detect and identify possible relations among the entities based on clustering without training data, as opposed to the relation extraction approach that focuses on improvement of accuracy in selecting one of the several target relations for a given pair of entities. While the relation extraction approach with supervised learning requires a significant amount of annotation efforts for a predefined set of relations, our approach attempts to discover relations as they occur naturally. Unlike other unsupervised relation identification work where evaluation of automatically identified relations is done with the correct relations determined a priori by human judges, we attempted to evaluate appropriateness of the naturally occurring clusters of relations involving person-artifact and person-organization entities and their relation names.

• Environmental Engineering Research
• /
• 제14권4호
• /
• pp.250-254
• /
• 2009

Contamination of microcystins, a family of heptapeptide hepatotoxins, in eutrophic water bodies is a worldwide problem. Due to their poisoning effects on animals and humans, there is a requirement to characterize and quantify all microcystins present in a sample. As microcystins are, for most part, intracellular toxins produced by some genera of cyanobacteria, lysing cyanobacterial cells to release all microcystins is considered an important step. To date, although many cell lysis methods have been used, little work has been conducted comparing the results of those different methods. In this study, various methods for cell lysis and toxin extraction from the cell lysates were investigated, including sonication, bead beating, freeze/thaw, lyophilization and lysing with TritonX-100 surfactant. It was found that lyophilization, followed by extraction with 75% methanol, was the most effective for extracting toxins from Microcystis aeruginosa cells. Another important step prior to the analysis is removing impurities and concentrating the target analyte. For these purposes, a C18 Sep-Pak solid phase extraction cartridge was used, with the percentage of the eluent methanol also evaluated. As a result, methanol percentages higher than 75% appeared to be the best eluting solvent in terms of microcystin-leucine-arginine (MC-LR) recovery efficiency for the further chromatographic and mass spectrometric analyses.

• 대한방사성의약품학회지
• /
• 제2권2호
• /
• pp.84-95
• /
• 2016

Considerably increasing interest in using the theranostic isotopes/ isotope pairs of radiometals like $^{44/47}Sc$ and $^{64/67}Cu$ for diagnosis and/or therapeutic applications in the nuclear medicine procedures necessitates its reliable production and supply. Separation and purification of no-carrier-added (NCA) isotopes from macro quantitates of the irradiated target matrix along with other impurities is a cardinal procedure amongst several other steps involved in its production. Multitudinous methods including but not limited to liquid-liquid (solvent) extraction, extraction chromatography (EXC), ion exchange, electrodeposition and sublimation are routinely applied either solitarily or in combination for the separation and purification of radioisotopes depending on their production routes, radioisotope of interest and impurities involved. However, application of EXC though has shown promises towards the numerous separation techniques have not received much attention as far as its application prospects in the field of nuclear medicine are concerned. Advances in the recent past for application of the EXC resins in separation and purification of the several medically important radioisotopes at ultra-high purity have shown promising behavior with respect to their operation simplicity, acidic and radiolytic stability, separation efficiencies and speedy procedures with the enhanced and excellent extraction abilities. In this mini review we will be talking about the recent developments in the application and the use of EXC techniques for the separation and purification of $^{44/47}Sc$ and $^{64/67}Cu$ for medical applications. Furthermore, we will also discuss the scientific and practical aspects of EXC in the view of separation of the NCA trace amount of radionuclides.

• 한국환경농학회지
• /
• 제25권3호
• /
• pp.236-246
• /
• 2006

In this study, we give insight into questionable results that can be encountered in the conventional sequential extraction of heavy metals (Cd, Cu, and Zn) from soils. Objectives of this study were to determine the extraction variability of exchangeable (EXC)-metals as using six different EXC-extractants commonly accepted, and to investigate selectivity problems with carbonates bound (CAB)-metal fraction, a buffered acetate (1.0 M NaOAc; pH 5.0) extractable-metal fraction, leading to erratic results in especially non-calcareous soils. The contents of EXC-metals were markedly varied with the different extractability of various EXC-metal extractants used. The contents of EXC-Cd fraction were ranged from 2.0 to 74.3% of total Cd content in all of the metal spiked soils studied. The contents of EXC-Zn fraction extracted with the different EXC-extractants were varied with soil types, which were from 0.4 to 3.9% of total Zn in the calcareous soils, from 7.6 to 17.9% in the acidic soil, and from 13.6 to 56.8% in the peat soil. However, the contents of EXC-Cu fraction were relatively similar among the applications of different EXC-meal extractants, 0.2 to 2.1 % of total Cu, in all soils tested. Also, these varied amounts of EXC-metal fractions, especially Cd and Zn, seriously impacted the contents of subsequent metal fractions in the procedure. Furthermore, the CAB-Cd, -Cu, and -Zn fractions extracted by the buffered acetate solution were in critical problem. That is, the buffered acetate solution dissolved not only CAB-metals but also metals that bound or occupied to subsequent fractions, especially OXD-metal fraction, in both calcareous and non-calcareous soils. The erratic results of CAB-fraction also seriously impacted the amounts of subsequent metal fractions. Therefore, the conventional sequential extraction should be reconsidered theoretically and experimentally to quantify the target metal fractions or might be progressively discarded.

• 한국IT서비스학회지
• /
• 제18권2호
• /
• pp.143-159
• /
• 2019

Machine learning (ML) is a method of fitting given data to a mathematical model to derive insights or to predict. In the age of big data, where the amount of available data increases exponentially due to the development of information technology and smart devices, ML shows high prediction performance due to pattern detection without bias. The feature engineering that generates the features that can explain the problem to be solved in the ML process has a great influence on the performance and its importance is continuously emphasized. Despite this importance, however, it is still considered a difficult task as it requires a thorough understanding of the domain characteristics as well as an understanding of source data and the iterative procedure. Therefore, we propose methods to apply deep learning for solving the complexity and difficulty of feature extraction and improving the performance of ML model. Unlike other techniques, the most common reason for the superior performance of deep learning techniques in complex unstructured data processing is that it is possible to extract features from the source data itself. In order to apply these advantages to the business problems, we propose deep learning based methods that can automatically extract features from transaction data or directly predict and classify target variables. In particular, we applied techniques that show high performance in existing text processing based on the structural similarity between transaction data and text data. And we also verified the suitability of each method according to the characteristics of transaction data. Through our study, it is possible not only to search for the possibility of automated feature extraction but also to obtain a benchmark model that shows a certain level of performance before performing the feature extraction task by a human. In addition, it is expected that it will be able to provide guidelines for choosing a suitable deep learning model based on the business problem and the data characteristics.

• 한국전자파학회논문지
• /
• 제23권8호
• /
• pp.1002-1009
• /
• 2012

본 연구는 미시간 주립대(Michigan State University)의 바이스태틱 레이다 시스템을 통하여 수집한 측정 데이터를 이용한 표적 구분에 관한 연구 결과이다. 본 연구에서는 먼저 F-14, Mig-29, F-22 스케일 모델에 대하여 $30^{\circ}$, $60^{\circ}$, $90^{\circ}$ 바이스태틱 각도에서의 측정을 수행하였다. 측정한 데이터로부터 시간-주파수 영역 해석법인 단시간 퓨리에 변환(Short Time Fourier Transform)과 연속 웨이브릿 변환(Continous Wavelet Transform)을 이용하여 특성 벡터를 추출하고, 신경망 구분기를 통하여 표적 구분 실험을 수행하였다. 실험 결과, 바이스태틱 각도에 따라 표적 구분 성능에 많은 변화가 있으며, 특히, $60^{\circ}$ 바이스태틱 각도에서 가장 좋은 구분 성능을 가짐을 알 수 있었다.

• 한국통신학회논문지
• /
• 제24권12B호
• /
• pp.2359-2368
• /
• 1999

본 논문에서는 2차원 영상을 기반으로 3차원 목표물을 인식하는 기법의 한 예로서 적외선 영상으로부터 추출된 물체의 모양 정보와 모양 정보의 신뢰도를 이용해서 지상에서 지상용 차량을 인식하는 기법(ground-to-ground vehicle recognition)을 제안한다. 우선 목표물 추출과정에서 얻어진 마스크의 윤곽선 상에 있는 점들 중 에지 경사도의 크기와 밝기값이 일정한 값 이상이 되는 점들을 신뢰도가 높은 점이라고 정의하고 신뢰도가 높은 점들을 연결해서 신뢰도가 높은 부분 윤곽선(sub-contour)을 추출한다. 모델로부터 입력 영상의 신뢰도가 높은 윤곽선에 해당되는 윤곽선을 선택한 후 각각 해당되는 윤곽선들은 이산 정현 변환(Discrete Sine Transform)을 사용해서 특징값을 계산한 다음 서로 비교한다. 실험 결과 영상 분할이 불완전한 경우 신뢰도를 이용한 방법이 그렇지 않은 방법보다 더 나은 결과를 보였다.

• 한국산학기술학회논문지
• /
• 제20권4호
• /
• pp.43-49
• /
• 2019

초고속의 유무선 인터넷은 다양한 형태의 미디어 데이터를 손쉽게 획득할 수 있는 유용한 창구이다. 이에 반해, 일반인들이 개인 정보가 노출된 대상 객체를 포함하고 있는 미디어 데이터까지도 인터넷을 통해 용이하게 획득할 수 있으므로 사회적으로 문제가 되고 있다. 본 논문에서는 입력되는 여러 가지 종류의 영상으로부터 개인 정보가 노출된 대상 객체를 학습 알고리즘을 이용해 강인하게 검출하고, 검출된 대상 객체 영역을 효과적으로 블로킹하는 방법을 제안한다. 본 논문에서 제안된 방법에서는 먼저 뉴럴 네크워크 기반의 학습 알고리즘을 사용해 영상으로부터 개인 정보를 포함하고 있는 대상 객체만을 검출한다. 그런 다음, 격자형 모자이크를 생성해 이전 단계에서 검출된 대상 객체 영역 위에 오버랩함으로써 개인 정보를 포함하고 있는 객체 영역을 효과적으로 블로킹한다. 실험 결과에서는 제안된 알고리즘이 입력되는 다양한 영상으로부터 개인 정보가 노출된 대상 영역을 강인하게 검출하고, 검출된 영역을 모자이크 처리를 통해 효과적으로 블로킹한다는 것을 보여준다. 본 논문에서 제시된 객체 블로킹 방법은 객체 보안, 물체 추적, 영상 블로킹 등과 같은 컴퓨터 비전과 관련된 여러 응용 분야에서 유용하게 활용될 것으로 예상된다.

• 정보처리학회논문지B
• /
• 제9B권2호
• /
• pp.245-254
• /
• 2002

머신비전을 이용한 IC 패키지 마킹검사 시스템은 입력영상으로부터 검사할 요소들의 위치를 식별하고, 추출된 요소들을 학습된 표준 패턴과 비교하여 마킹의 불량 여부를 판단한다. 본 논문에서는 검사 대상 IC 패키지의 위치 판별, 마킹문자 추출, 핀원딤플 검출과 같은 일련의 작업들에 적합한 적응적 다단계 이진화 방법과 마킹문자의 국소적인 오류검출은 물론 잡영에 강건한 정합단위의 동적 선택 방법을 제안한다. 제안하는 이진화 방법은 이진화 대상 영역과 명도 값의 범위를 제한하여 Otsu의 이진화 알고리즘을 적용함으로써 특정 응용에 적응적인 이진화가 가능하다. 정합단위의 동적 선택 방법은 문자추출 및 배치분석에 대한 결과에 따라 정합단위를 선택한다. 그러므로 문자추출 및 배치분석 과정에서 발생하는 예기치 못한 부적절한 상황에서도 가능한 범위내에서 최소의 정합단위를 선택할 수 있다. 제안된 방법을 구현하여 8종의 IC 패키지, 총 280개의 영상에 대하여 실험한 결과, IC 패키지와 핀원딤플의 검출율은 100%였으며, 마킹상태에 대한 판정은 98.8%의 정확도를 나타내어 제안된 방법이 효과적임을 확인할 수 있었다.

• Bulletin of the Korean Chemical Society
• /
• 제35권9호
• /
• pp.2665-2668
• /
• 2014

Aptamers are oligonucleotides or peptide molecules that are able to bind to their specific target molecules with high affinity via molecular recognition. In this study, we present development of aptamer-based precipitation assays (or simply aptamoprecipitation) for His-tagged proteins and thrombin to compare their purification efficiency with other conventional affinity precipitation methods. A crosslinking method was employed to immobilize thiol-functionalized aptamers onto the surface of polystyrene resins, enabling them to specifically bind to His-tag and to thrombin, respectively. The resulting aptamer-functionalized resins were successfully applied via a one-step experiment to purification of His-tagged proteins from complex E. coli and to thrombin extraction, exhibiting superior or at least comparable purification results to the conventional immobilized metal affinity precipitation or immunoprecipitation.