• Journal of the Mineralogical Society of Korea
• /
• v.23 no.1
• /
• pp.85-91
• /
• 2010

The various forms of heavy metals in soil environments have been studied by sequential extraction method. We tested conventional Tessier sequential extraction and new ultrasound-sonication extraction methods, and compared their extraction efficiency. Total As, Cd, Cu, Pb, and Zn contents of the target soil (NIST SRM 2710 Montana Soil), by three methods (USEPA Method 3050B, KBSI Method, and ultrasound-sonication method) were all consistent with the certified values. Sequential extraction efficiency along with step-wise extraction values was similar in both Tessier method and ultrasound-sonication method. The ultrasound-sonication method took about 3 hours to complete whole procedure while the Tessier sequential extraction method took around 12 hours. Ultrasound-sonication method was estimated as one of the best methods with a relatively short application time and no requirement of high temperature treatment.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06c
• /
• pp.462-467
• /
• 2008

최근 웹 2.0 시대 참여, 공유, 개방 정신이 확대 되고, 다양한 디지털 저작물들이 대량 제작되어 활용되고 있다. 그리고 디지털 저작물의 특징상 누구나 손쉽게 무제한으로 복제와 유통이 가능함으로 디지털 저작물이 양은 기하급수적으로 증가하고 있다. 증가하는 이미지를 효과적으로 관리하고 검색하기 위해 색상, 질감, 모양 등을 이용한 내용기반 이미지 검색에 대한 연구가 활발히 진행되고 있다. 색상을 이용한 이미지 검색방법의 하나로 색상 히스토그램을 이용한 검색 방법이 있다. 그러나 이는 공간적인 상호관계를 적절히 표현하지 못한다는 단점이 있다. 따라서 본 논문에서는 이미지에 나타나는 주요 색상 및 불변 모멘트 값과 이미지의 중앙을 중심으로 한 영역별 유사도 검사를 통한 내용기반 이미지 검색 시스템을 제안한다. 첫 번째 유사성 검사 단계에서는 이미지의 영역별로 가중치를 부여하여 추출한 대표색상 정보를 사용하여, 유사하지 않은 이미지를 제거하여 검색대상의 수를 줄인다. 두 번째 유사성 검사 단계에서는 이미지를 영역으로 나누고, 이미지의 중심 영역부터 영역을 확장하며 영역마다 구축된 인덱스 검색을 통해 영역기반 유사 이미지 검색을 수행 한다. 세 번 단계에서는 이미지의 변형에 불변한 값인 불변 모멘트를 사용하여, 영역별 검사에서 제외된 유사이미지를 재검사한다. 제안한 이미지 검색 방법은 10000개의 다양한 이미지로 구성된 이미지 데이터베이스에서 검색을 실험을 통해 검색의 정확도 및 회수율을 측정하였다.

• Journal of IKEEE
• /
• v.13 no.4
• /
• pp.1-7
• /
• 2009

A video segmentation algorithm, which can extract objects even with non-stationary backgrounds, is proposed in this work. The proposed algorithm is composed of three steps. First, we perform an initial segmentation interactively to build the probability density functions of colors per each macro block via kernel density estimation. Then, for each subsequent frame, we construct a coherence strip, which is likely to contain the object contour, by exploiting spatio-temporal correlations. Finally, we perform the segmentation by minimizing an energy function composed of color, coherence, and smoothness terms. Experimental results on various test sequences show that the proposed algorithm provides accurate segmentation results.

• Korean Journal of Cognitive Science
• /
• v.19 no.4
• /
• pp.443-458
• /
• 2008

Since Web pages contain a lot of information today, the Web becomes an important resource to extract some information. In this paper, we proposes a set expansion system which can automatically extract named entities from the Web. Overall, the proposed method consists of three steps. First of all, Web pages, which may include many named entities of a domain, are collected by using several seed words of the domain. Then some pattern rules are extracted by using seed words and the collected Web pages, and the named entity candidates are selected through applying the extracted pattern rules into Web pages. To distinguish real named entities, we develop the new mutual importance measurement method which estimates the importance of named entity candidates. We conducted experiments for 3 domains for Korean and for 8 domains for English. As a result, the proposed method obtained 78.72% MAP in Korean and 96.48% MAP in English. In particular, the performances of English domains are better than the results of the Google set.

• Annual Conference on Human and Language Technology
• /
• 2003.10d
• /
• pp.89-94
• /
• 2003

본 논문에서는 질의문 내에 포함된 동사를 중심으로 한 질의어 확장 및 정답 추출 기법을 이용한 질의 응답 시스템에 대해 기술한다. 질의 응답시스템 전체의 과정에서 동사는 하나의 정보를 표현하는 중요한 요소로 활용하며, 동사에 대한 활용은 구축된 동사구문 사전의 정보를 이용한다. 동사구문 사전은 동사의 일반적인 표층형태와, 각 문장 성분들의 의미속성, 유의동사 등의 정보를 담고 있다. 또한 동사 구문사전의 활용에서의 동사 모호성을 배제하고, 효율을 높이기 위해 약 3만 어휘의 명사 의미 사전을 사용한다. 명사 의미사전은 구문사전 내에 사용된 의미분류로 나누어져 있으며, 유의명사 및 국어사전 상의 뜻 풀이말을 포함하고 있다. 질의문 및 각 후보 문장에 대한 구문분석은 구문사전 내에 나타난 품사 별 의미속성과, 문법 형태소의 격 정보를 이용한 격 구조를 활용하였다. 논문 중에는 일반적인 질의 응답 시스템의 3단계에 맞추어 구문사전 활용 및 구문분석의 수행 단계를 보이고 마지막에 각 기법의 정확도를 보였다.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.5
• /
• pp.69-76
• /
• 2022

In this paper, we presented the contents and methods of AI classes using AI platforms. First, we extracted the content elements of each stage of the AI class using the AI education platform from experts. Classes using the AI education platform were divided into 5 stages and 25 class elements were selected. We also conducted a survey of 82 teachers and analyzed the factors that they acted importantly at each stage of the AI platform class. As a result of the analysis, teachers regarded the following contents as important factors for each stage that are AI model preparation stage (the learning stage of the AI model), problem recognition stage (identification of problems and AI solution potential), data processing stage (understanding the types of data), AI modelingstage (AI value and ethics), and problem solvingstage (AI utilization in real life).

• The Journal of the Korea Contents Association
• /
• v.19 no.1
• /
• pp.125-131
• /
• 2019

A vanishing point is a point on an image to which parallel lines projected from a real space gather. A vanishing point in a road space provides important spatial information. It is possible to improve the position of an extracted lane or generate a depth map image using a vanishing point in the road space. In this paper, we propose a method of detecting vanishing points on images taken from a vehicle's point of view using Deep Neural Network (DNN) and Histogram of Oriented Gradient (HoG). The proposed algorithm is divided into a HoG feature extraction step, in which the edge direction is extracted by dividing an image into blocks, a DNN learning step, and a test step. In the learning stage, learning is performed using 2,300 road images taken from a vehicle's point of views. In the test phase, the efficiency of the proposed algorithm using the Normalized Euclidean Distance (NormDist) method is measured.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.251-254
• /
• 2003

데이터는 어떤 특성을 나타내는 언어적 또는 수치적 값들의 표현이다. 이러한 데이터들을 목적에 따라 구성한 것이 정보이며, 문제 해결이나 패턴 분류, 또는 의사 결정을 위해 정보들간의 관계를 규칙으로 체계화하는 것이 지식이다. 현재 대부분의 산업 분야에서 시스템에 대한 이해를 높이고 시스템의 성능을 향상시키기 위해 지식을 추출하고, 적용시키는 작업들이 활발히 이루어지고 있다. 지식 정보의 추출은 지식의 획득, 표현, 구현의 단계로 구성되며 이렇게 추출된 지식 정보는 규칙으로 도출된다. 본 논문에서는 여러 산업 분야에 걸쳐 다양하게 적용되는 지식 정보 추출 방법들에 대해 그 영역별로 알아보고 여러 시험 데이터들과 실제 시스템에 클러스터링(CL), 입력공간 분할(ISP), 뉴로-퍼지(NF), 신경망(NN), 확장 행렬(EM) 등의 방법들을 적용시킨 결과들을 비교 분석하고자 한다.

• Korean Journal of Remote Sensing
• /
• v.19 no.4
• /
• pp.307-317
• /
• 2003

This paper presents an algorithm that automatically extracts buildings among many different features on the earth surface by fusing LIDAR data with panchromatic aerial images. The proposed algorithm consists of three stages such as point level process, polygon level process, parameter space level process. At the first stage, we eliminate gross errors and apply a local maxima filter to detect building candidate points from the raw laser scanning data. After then, a grouping procedure is performed for segmenting raw LIDAR data and the segmented LIDAR data is polygonized by the encasing polygon algorithm developed in the research. At the second stage, we eliminate non-building polygons using several constraints such as area and circularity. At the last stage, all the polygons generated at the second stage are projected onto the aerial stereo images through collinearity condition equations. Finally, we fuse the projected encasing polygons with edges detected by image processing for refining the building segments. The experimental results showed that the RMSEs of building corners in X, Y and Z were 8.1cm, 24.7cm, 35.9cm, respectively.

• Journal of Life Science
• /
• v.18 no.8
• /
• pp.1123-1131
• /
• 2008

This study was performed to confirm of biological function of black garlic, it heated for 11 days at $40{\sim}90^{\circ}C$ (first step: heate for 2 days at $90^{\circ}C$, second step: heated for 4 days at $80^{\circ}C$, third step: heated for 4 days at $60^{\circ}C$ and fourth step: heated for 1 day at $40^{\circ}C$). Samples were analyzed physico-chemical characteristics and antioxidant activity. Hunter L, a and b values were decreased during processing, and then inner part Hunter values were highly decreased at the second step. The moisture contents were decreased to 58.48${\pm}$0.41 g/100 g at fourth step. pH was also acidified to pH 4.22${\pm}$0.02, but O.D. value at 420 nm was increased during processing of black garlic. At fourth step, total phenolics and flavonoids contents were increased about 1.9 and 2.6 folds than first step sample. Also, total pyruvate and thiosulfinate contents were increased about 1.6 and 5.8 folds as change of total phenolic and flavonoid contents, respectively. Fructose contents were the highest level among free sugars and its contents increased to 2,454.45${\pm}$4.20 mg/100 g. Contents of sucrose and maltose were decreased during processing of black garlic. The contents of total minerals were the highest at fourth step (1,009.20${\pm}$6.91 mg/100 g) during its processing. Contents of glutamic acid, proline and aspartic acid were detected higher than other composition amino acids. Taurine and ethanolamine were not detected in the first step sample, but they were detected 0.88${\pm}0.60{\sim}1.06{\pm}$0.04 and 0.28${\pm}0.4{\sim}0.5{\pm}$0.09 mg/100 g in next processing step, respectively. DPPH radical scavenging ability of water and ethanol extracts from black garlic was increased during its processing. Abilities of DPPH radical scavenging were the highest in fourth step sample, its abilities were 67.40${\pm}$0.21% in 1,000 ${\mu}g/ml$ of water extracts. Reducing power was also significantly higher in water extract than ethanol extract on the whole.