• 터널과지하공간
• /
• 제15권1호
• /
• pp.39-46
• /
• 2005

터널 단층대에서 수평시추로 조사한 막장전방의 암반 상태를 공학적 암반분류법인 RMR값으로 평가하였고 이를 터널 굴착 후 막장을 관찰하여 결정한 RMR값과 비교 분석하였다. 수평시추로 예측한 RMR값은 비교적 정확하여 터널 굴진 후 막장을 관찰하여 구한 RMR값과 큰 차이가 없었다. 그러나 일부 구간에서는 수평시추와 막장관찰로 구한 RMR값의 차이가 약 50까지 발생하였고 이를 RMR 평가항목으로 분석한 결과 불연속면의 상태에 대한 평점에서 24의 차이가 나타났고 암질지수와 단축압축강도 평점에서 각각 15와 13의 차이로 나타났다. 두 방법에서 평가한 RMR값의 차이를 줄이기 위해서는 터널 내 수평시추공의 위치를 터널의 안정성에 가장 큰 영향을 줄 수 있는 곳으로 선정하고 불연속면의 상태에 대한 평가는 불연속면의 연속성, 분리 틈, 풍화도 등 5개의 소항목 각각에 대해 5단계로 구분한 세부평점을 적용하여야 할 것이다.

• 한국산학기술학회논문지
• /
• 제21권8호
• /
• pp.395-400
• /
• 2020

전 세계적으로 유행하며 수많은 확진자와 사망자를 발생시킨 코로나바이러스-19(COVID-19)는 일상에서 사람 간 전염이 가능하여 국민들을 불안과 공포에 떨게 하고 있다. 감염을 최소화하기 위해서는 건물 출입시 마스크 착용이 필수적이지만 일부 사람들은 여전히 마스크 없이 얼굴을 노출시킨 채 건물에 출입하고 있다. 본 논문에서는 효율적인 출입 통제를 위해 얼굴에 마스크를 착용했는지 여부를 자동으로 판별하는 방법을 제안한다. 제안 방법은 양쪽 눈 영역을 검출하고 눈 위치를 참조하여 마스크 착용 영역(양쪽 눈 아래 얼굴 영역)을 예측한다. 이 때 마스크 착용 영역을 보다 정확히 예측하기 위해 양쪽 눈 위치가 수평이 되도록 얼굴 영역을 회전하여 정렬한다. 정렬된 얼굴 영역에서 추출된 마스크 착용 영역은 이미지 분석에 특화된 딥러닝 기법인 CNN(Convolutional neural network)을 통해 마스크 착용 여부(착용 또는 미착용)를 최종 판별한다. 총 186장의 테스트 이미지에 대해 실험한 결과, 98.4%의 판별 정확도를 보였다.

• 사상체질의학회지
• /
• 제23권2호
• /
• pp.194-207
• /
• 2011

1. Objectives: This study is about a development of Sasang constitutional classification algorithm using facial information. 2. Methods: We analysed the datum of middle aged (20~48) women collected by multi-center researchers in 2007. And this study analysed the data of the measurement of the face by 3D-AFRA (3-Dimensional Automatic Face Recognition Apparatus) and the items of impression by SDQ. We used multiple comparison, exploratory discriminant analysis and clinical decision to select optimal 3D facial variables which will be input in discriminant analysis model. And we used univariate F values and stepwise discriminant function analysis to choose best impression variables. 3. Results and Conclusions: In this study, derived discriminant function's explanation power was 39% in female group. Diagnostic accuracy rate was 66.0% in female group. And in test sample, Sasang constitutional diagnostic accuracy rate was 56.9%. In this process we could help improve the objectification of Sasang constitution diagnosis.

• 한국IT서비스학회지
• /
• 제20권5호
• /
• pp.137-146
• /
• 2021

Recently, public participation in government policy design has been further expanded and public services perceived by users are expanding. At this time, the role of the digital government and the direction of the service to be pursued are user-centered, and above all, it is necessary to focus on the keywords of pre-emptive, preventive, and customized. In order to propose service quality improvement in the public sector, service user-centered classification and monitoring are integrated and the usability of government documents is improved. It is necessary to identify the needs of whether to provide a path for public participation. In the post-corona era, people are accessing quarantine information from the digital government every day. The government should proactively respond to the acceleration of digital transformation and the non-face-to-face demands of the people who experience non-face-to-face daily life. In order to evolve into a smart organization along with the innovation promotion plan and to provide customized services, it is necessary to use existing guides for institutional and technical improvement, along with new technology and data-based analysis, to strive for change management. The government should seek counter-measures that have advanced one step ahead by incorporating new high-tech IT with user-centered necessary services. This study aims to derive improvement plans to provide user-centered digital government service design when designing public services and collecting public opinions. Based on the e-government development model research and the existing research on user-centered service design in the public sector, institutional and technical measures are provided for the improvement of digital government service design.

• Genomics & Informatics
• /
• 제2권3호
• /
• pp.121-125
• /
• 2004

The purpose of this study was to investigate the use of decision tree for the classification of antimicrobial peptides. The classification was based on the activities of known antimicrobial peptides against common microbes including Escherichia coli and Staphylococcus aureus. A feature selection was employed to select an effective subset of features from available attribute sets. Sequential applications of decision tree with 17 nodes with 9 leaves and 13 nodes with 7 leaves provided the classification rates of $76.74\%$ and $74.66\%$ against E. coli and S. aureus, respectively. Angle subtended by positively charged face and the positive charge commonly gave higher accuracies in both E. coli and S. aureusdatasets. In this study, we describe a successful application of decision tree that provides the understanding of the effects of physicochemical characteristics of peptides on bacterial membrane.

• 한국의류학회지
• /
• 제28권7호
• /
• pp.983-994
• /
• 2004

The aim of this study is to provide fundamental data on the development of ready-to-wear clothes appropriate for the body types of elderly women. The study was conducted targeting 318 elderly women over 60 years of age whose fields of action were colleges for the elderly, sports centers, or business sites in Seoul and the neighboring districts. A total of 44 features in the upper body were used for the anthropometric measurement and analysis using anthropometry and photometry. The results of the study are as follows: 1. Somatotypes were classified into three types according to a cluster analysis using height and weight indices. Type 1 is the group with long and undersized upper body and straight body type since the face of the upper body is long relative to height and width, girth and depth are the smallest relative to weight, the breasts are somewhat fat, with a small extent of drooping and a straight back. Type 2 is the group that is considered fat relative to the body, has broad shoulders, drooping breasts with a wide space between them, and a back-bent upper body. Type 3 is the group that has a bent shape, the shortest upper body relative to height, and showing average obesity factors. 2. Indices of height and weight were used for factor analysis, cluster analysis, and discriminant analysis in order to classify upper body somatotype according to shape while excluding size factors of elderly women＇s upper body somatotype. The same method was used to compare and verify the result according to the absolute measurement and height index. Classification based on height and weight indices demonstrate that such somatotype classification minimizes the personal equation of body shape and it induces better classification based on shape as the results showed the highest cumulative sum of square(CUSUM) at 78.38% while six factors showed the smallest result and the hit rate for the classified three groups showed the highest result at 95.30%.

• 대한전자공학회논문지SP
• /
• 제49권2호
• /
• pp.102-112
• /
• 2012

본 논문에서는 웨이브렛 영역의 BDIP(block difference of inverse probabilities)와 BVLC(block variance of local correlation coefficients) 특징, 그리고 WPCA(whitened principal component analysis) 분류기를 이용한 질감 분류 방법을 제안한다. 제안된 방법에서는 먼저 질의 영상에 웨이브렛 변환을 적용한다. 그런 다음 웨이브렛 영역의 각 부대역에 BDIP와 BVLC 연산자를 적용한다. 이어서 각 BDIP, BVLC 부대역에 대하여 전역 통계치를 계산하고 그 결과들을 벡터화하여 특징 벡터로 사용한다. 분류 단계에서는 얼굴 인식에 주로 사용되는 WPCA를 분류기로 하여 질의 특징 벡터와 가장 유사한 학습 특징 벡터를 찾는다. 실험 결과 제안된 방법은 3가지의 실험 질감 영상 DB에 대하여 낮은 특징 벡터 차원으로 매우 우수한 질감 분류 성능을 보여준다.

• International Journal of Computer Science & Network Security
• /
• 제22권2호
• /
• pp.272-282
• /
• 2022

Between 2014 and 2019, the US lost more than 2.1 billion USD to phishing attacks, according to the FBI's Internet Crime Complaint Center, and COVID-19 scam complaints totaled more than 1,200. Phishing attacks reflect these awful effects. Phishing websites (PWs) detection appear in the literature. Previous methods included maintaining a centralized blacklist that is manually updated, but newly created pseudonyms cannot be detected. Several recent studies utilized supervised machine learning (SML) algorithms and schemes to manipulate the PWs detection problem. URL extraction-based algorithms and schemes. These studies demonstrate that some classification algorithms are more effective on different data sets. However, for the phishing site detection problem, no widely known classifier has been developed. This study is aimed at identifying the features and schemes of SML that work best in the face of PWs across all publicly available phishing data sets. The Scikit Learn library has eight widely used classification algorithms configured for assessment on the public phishing datasets. Eight was tested. Later, classification algorithms were used to measure accuracy on three different datasets for statistically significant differences, along with the Welch t-test. Assemblies and neural networks outclass classical algorithms in this study. On three publicly accessible phishing datasets, eight traditional SML algorithms were evaluated, and the results were calculated in terms of classification accuracy and classifier ranking as shown in tables 4 and 8. Eventually, on severely unbalanced datasets, classifiers that obtained higher than 99.0 percent classification accuracy. Finally, the results show that this could also be adapted and outperforms conventional techniques with good precision.

• 한국컴퓨터정보학회논문지
• /
• 제26권5호
• /
• pp.39-46
• /
• 2021

인공지능은 현재 인공지능 번역기, 페이스 아이디와 같이 우리의 삶 다양한 곳에 적용되고 있으며 여러 가지 장점으로 많은 산업분야에서도 적용되고 있다. 본 연구는 매년 방대한 양의 콘텐츠들이 넘쳐나는 상황에서 인공지능을 적용한 카테고리 분류로 원하는 데이터를 추출함으로써 편의성을 제공한다. 본 연구에서는 텍스트 분류에서 두각을 나타내고 있는 LSTM(Long-Short Term Memory network)을 사용한 모델을 제안하며 자연어 처리에 적합한 구조를 가진 RNN(Recurrent Neural Network)과 BiLSTM(Bidirectional LSTM)을 사용한 모델과의 성능을 비교한다. 세 가지 모델의 성능비교는 뉴스 텍스트 데이터에 적용해 accuracy, precision, recall의 측정값을 사용해 비교하였고 그 결과 LSTM모델의 성능이 가장 우수한 것으로 나타났다. 따라서 본 연구에서는 LSTM을 사용한 텍스트 분류를 권장한다.

• 사상체질의학회지
• /
• 제18권2호
• /
• pp.41-49
• /
• 2006

1. Objectives The Face is an important standard for the classification of Sasang Contitutions. We are developing 3D Face Automatic Recognition Apparatus to analyse the facial characteristics. So We should examine a shape demobilization error of 3D Face Automatic Recognition Apparatus. 2. Methods We compared facial shape data be demobilized by 3D Face Automatic Recognition Apparatus with facial shape data that be demobilized by 3D laser scanner. The subject was two korean men. And We analysed the average error and the maximum error of two data. In this process, We used one datum point(the peak of nose) and two datum line(vertical section and horizontal section). 3. Results and Conclusions In each this comparison, the average error of vertical section was 1.962574mm and 2.703814mm. and the maximum error of vertical section was 16.968249mm and 18.61464mm. the average error of horizontal section was 4.173203mm and 21.487479mm. and the maximum error of horizontal section was 3.571210mm and 17.13255mm. Also We complemented this apparatus a little and We reexamined a shape demobilization error of 3D Face Automatic Recognition Apparatus again. Accuracy of a shape demobilization was improved a little. From now on We complement accuracy of a shape demobilization in 3D Face Recognition Apparatus.