무인기와 주변 개체간의 위협 관련 관계 추론을 통한 무인기 상황인지 기법

배석현,전명중,박현규,박영택,윤형식,김윤근 한국정보과학회 2019 정보과학회논문지 Vol.46 No.2

As the technological capabilities of UAV(Unmanned Aerial Vehicles) improves, studies are being carried out to intelligently analyze and understand the situation of UAV in order to gain access to the target area while recognizing and avoiding various risks. To achieve the mission of UAV, it is necessary to judge the situation accurately and quickly. To do this, this paper proposes ways to infer the threat-related relationship between an UAV and perceived surrounding objects through a 3 step approach and provide abstract information about the situation of UAV. The first step is to instantiate the object data recognized by UAV to be utilized for ontology and rule-based reasoning. The second step is to define the priority of instantiated object data and to infer the threat-related relationship between them. Finally, recognizing the situation through the relationship inference that takes into account the association between current and past inferred relationships. To evaluation the performance of the proposed method, a virtual UAV environment simulator was built and tested the data 1,000 times that were randomly generated through five sequential UAV moving point paths. Eight kinds of objects could be recognized in UAV path and ten kinds of relationships can be inferred. Overall performance of situation Awareness was an average of 91 percent. 무인기의 기술적 성능이 향상됨에 따라 무인기가 각종 위험을 스스로 인지하고 회피하면서 목표 지역에 접근하기 위해 지능적으로 UAV의 상황을 분석하고 이해하려는 연구가 활발하다. 무인기의 임무를 달성하기 위해서는 빠른 상황 판단과 함께 정확한 상황 판단이 요구된다. 이를 위해 본 논문에서는 3단계의 접근방식을 통하여 무인기와 인지된 주변 개체 간의 위협 관련 관계를 추론하고 이를 기반으로 무인기의 상황을 추상화된 정보로 제공하는 방법을 제안한다. 첫 번째 단계는 무인기가 인지한 개체 데이터를 온톨로지 및 규칙 추론에 활용하기 위해 개체화 하는 것이다. 두 번째 단계는 개체화된 데이터에 대해 개체 간 위협과 관련된 추론의 우선순위를 정의하고 이들 간의 관계 추론을 한다. 마지막으로 현재 추론된 관계들과 과거에 추론된 관계들 간의 연관성을 고려한 관계 추론을 통하여 상황을 인지한다. 제안한 방식의 성능 평가를 위해 가상의 무인기 환경 시뮬레이터를 구축하고 순차적인 5개의 무인기 이동 포인트 경로를 무작위로 1,000번 생성하여 실험하였다. 무인기 이동 경로에서 8종류의 개체를 인지할 수 있으며, 10종류의 관계를 추론할 수 있다. 그리고 전체적인 추론 인지 성능은 평균 91% 이다.

돈피의 입자 크기에 따른 유화소시지의 품질특성 변화

배석현,조병욱,조성근,김병우,서자겸,유장곤,신택순 경상대학교 농업생명과학연구원 2018 농업생명과학연구 Vol.52 No.5

본 연구는 지방 대체제로 첨가된, 각기 다른 입자의 크기로 분쇄된 돈피가 유화소시지의 이화학적 성질과 저장 기간, 그리고 관능검사에 끼치는 영향을 조사하기 위해 수행되었다. 지방 대체제로 첨가된 돈피는 대조구에 첨가된 지방 12%와 동일한 수준인 지방 4%, 분쇄 돈피 8%를 T1에서는 3mm, T2에서는 5mm, T3에서는 7mm로 분쇄하여 첨가하였다. 모든 처리구들은 일반적인 냉장 온도인 4±1℃에서 0, 5, 10, 15일간 저장되며 시료로 쓰였다.첨가되는 돈피의 분쇄 입자 크기가 클수록 전단가와 경도, 검성, 씹힘성이 유의적으로 증가하였고, 전 저장구간에서 포장감량, TBARS, VBN, 씹힘성은가식권 내의 값을 유지하며 유의적으로 증가하였다. 관능검사에서는 Color에서 입자의 크기가 클수록 부정적인 영향이 있었고, Chewiness는 저장기간이 지남에 따라 유의적으로 증가하였고 입자 크기가 작을수록 유의적 으로 높은 값을 나타내었다. 실험결과 돈피를 첨가한 처리구가 대조구에 비해 Shear force, Puncture test, 관능 평가에서 우수한 평가를 받았고 저장성검사와 이화학적 성질에서도 큰 차이를 보이지 않고 모두 가식권내의 값을 유지하였다. 이는 돈피가 육가공품에서 지방이 작용하는 역할을 충분히 수행가 능하며, 좋은 평가를 얻어 지방을 대체할 수 있을 것으로 판단된다. The aim of this study was that how grounded pork rind effects on physico-chemical properties, storage period and sensory evaluation of emulsified sausage as fat substitute. There was total 12% of fat in control, and 8% of fat from treatments was replaced by grounded pork rind as fat substitute in a different sizes(3, 5, 7 mm). All the samples were stored 0, 5, 10, 15 days in conventional refrigerator temperature of 4±1℃. Bigger pork rind particle increased shear force, hardness, gumminess and chewiness significantly, purge loss, TBARS, VBN and chewiness increased significantly during the storage period. While supplementation of pork rind had negative effect on color, it increased chewiness significantly with smaller size and during the storage period. Consequently, treatments had higher score from shear force, puncture test and sensory evaluation, and remained eatable in storage and physico-chemical properties. In summary, it turned out that pork rind was good enough to work like fat so that it can act like fat substitute.

Current Status of Collaboration and Research Trend in Microbiology in South Korea

배석현,조근태 한국경영과학회 2018 經營 科學 Vol.35 No.4

The objective of the study is to identify how researchers conduct collaborative work and which is the trend in the field of Korean microbiology using scientometrics and social network analysis (SNA) approaches from Korea-based microbiology articles. For the purpose, we use co-authorship as an indicator of the research collaboration network by analyzing 7,756 published sources from representative Korean microbial articles between 1996 and 2015. The results show that cooperation with different institutions is gradually increasing over time but the proportion of co-authorship within an institution decreased. And the average number of authors per paper is 4.928 during the last two decades in the Korean Microbiology Journal. Moreover, microbiology research trends in Korean have shifted with the passage of time from microbes and enzymes themselves to genomes, recombinant methods, proteins, bio-engineering, immunity, and molecular diagnosis. The significance of this study is that it is the first attempt to analyze the trends in Korean microbiology research using scientometrics with representative microbiology articles published by Korean scholars and using co-occurrent keywords to identify the research trend of Korean microbiology.

Preoperative Embolization of a Giant Neurofibroma of the Chest in a Patient with Neurofibromatosis Type II: A Case Report

배석현,신종수 대한영상의학회 2017 대한영상의학회지 Vol.76 No.1

Giant plexiform neurofibromas, which are rare in patients with neurofibromatosis type II (NFII), are difficult to manage surgically, as they are extensively infiltrative and highly vascularized. Preoperative embolization is performed to reduce intraoperative blood loss and operative time, increase resectability of lesions, and improve visualization of the operative field during surgery of hypervascular tumors such as renal cell carcinoma and intracranial meningioma. Preoperative intravascular embolization of a giant chest wall neurofibroma has not been reported in the English literature. We report successful treatment of a giant chest wall neurofibroma in a 45-year-old male with NFII by preoperative intravascular embolization followed by surgical resection.

Influence of Myopia on Size of Optic Nerve Head and Retinal Nerve Fiber Layer Thickness Measured by Spectral Domain Optical Coherence Tomography

배석현,강신희,풍계현,박주현,정재훈,이가영 대한안과학회 2016 Korean Journal of Ophthalmology Vol.30 No.5

Purpose: To investigate optic nerve head size and retinal nerve fiber layer (RNFL) thickness according to refractivestatus and axial length. Methods: In a cross-sectional study, 252 eyes of 252 healthy volunteers underwent ocular biometry measurementas well as optic nerve head and RNFL imaging by spectral-domain optical coherence tomography. Correlationand linear regression analyses were performed for all subjects. The magnification effect was adjustedby the modified axial length method. Results: Disc area and spherical equivalent were positively correlated (r = 0.225, r2 = 0.051, p = 0.000). RNFLthickness showed significant correlations with spherical equivalent (r = 0.359, r2 = 0.129, p = 0.000), axiallength (r = -0.262, r2 = 0.069, p = 0.000), disc radius (r = 0.359, r2 = 0.129, p = 0.000), and radius of the scancircle (r = -0.262, r2 = 0.069, p = 0.000). After adjustment for the magnification effect, those relationships werereversed; RNFL thickness showed negative correlation with spherical equivalent and disc radius, and positivecorrelation with axial length and radius of the scan circle. The distance between the disc margin and the scancircle was closely correlated with RNFL thickness (r = -0.359, r2 = 0.129, p = 0.000), which showed a negativecorrelation with axial length (r = -0.262, r2 = 0.069, p = 0.000). Conclusions: Optic disc radius and RNFL thickness decreased in more severely myopic eyes, but they increasedafter adjustment for magnification effect. The error due to the magnification effect and optic nervehead size difference might be factors that should be considered when interpreting optical coherence tomographyresults.

참여 있는 곳에 단합이

裵錫鉉 대한사립학교장회(구 대한사립중고등학교장회) 1977 私學 Vol.2 No.-

돈피의 첨가 수준에 따른 유화소시지의 품질특성 변화

배석현,조병욱,조성근,김병우,서자겸,유장곤,신택순 경상대학교 농업생명과학연구원 2015 농업생명과학연구 Vol.49 No.5

본 실험은 돈피를 첨가하지 않은 대조구와 돈피를 4% 첨가한 처리구 1, 8% 첨가한 처리구 2, 12% 첨가한 처리구 3으로 돈피 첨가수준을 달리하여 제조된 유화소시지를 4±1℃의 냉장온도에 저장하면서 15일간 5일 간격으로 이화학적 성질, 저장성 실험, 관능검사를 실시하였다. 제품 제조시 돈피를 지방 대체제로 사용하여 돈피의 첨가 비율만큼의 지방은 제외하였다. 지방의 비율이 낮아지고 돈피 첨가수준 이 높아짐에 따라 pH, 전단가, 경도, 검성와 씹힘성은 전 처리구에서 유의적으로 증가했고, 지방산패도 (TBARS: thiobarbituric acid reactive substances)와 휘발성 염기태질소(VBN: volatile basic nitrogen)도 유의적 차이는 나타나지 않았고 증가하는 경향을 보였지만 전 저장기간 동안 가식권에 있 었다. 관능평가에 있어서는 돈피를 8%를 첨가한 처리구 2가 가장 좋은 평가를 받았다. 저장기간이 경 과함에 따라 포장 감량(purge loss), 연도, 경도, 검성, 씹힘성은 전 처리구에서 유의적으로 증가하였으 나(p<0.05), TBARS와 VBN은 증가하였으나 유의성을 나타내지 않았다. pH는 대조구에서는 증가했으나 나머지 처리구에서는 변화가 없었다. 이상의 결과를 종합하면, 대조구와 전 처리구에서 이화학적 성질 이 가식권 내 값을 유지하므로 관능평가에서 잠재적인 소비자들에게 우수한 기호성을 획득한 돈피 함량 이 8%가 최적의 첨가량임을 알 수 있다. This research was designed to investigate changes in quality based on physico-chemical properties and sensory quality of emulsified sausages with different levels of pork rind added during different storage periods of 0, 5, 10 and 15 days. As pork rind is added as a replacer for fat in emulsified sausage, the same ratio of fat is removed from the samples. The research was conducted with four groups of emulsified sausages including different pork rind levels(Control: 0%, T1: 4%, T2: 8%, T3: 12% pork rind added) and stored at conventional refrigerator at 4±1℃. As more amount of pork rind was added, all of the physico-chemical values of emulsified sausages not only pH, shear force, hardness, gumminess and chewiness were increased, but TBARS and VBN were also increased. Although there was no great differences among all of the values of physico-chemical properties from four treatments, T2(8%) got the best score in sensory quality. All of the values of physico-chemical properties are increased according to the storage periods within safety level, including the values of pH of C. Based on the results in this study, adding 8% of pork rind as replacer of fat is the best for emulsified sausages, reducing the certain amount of fat.

An Approach to a Learning Prediction Model for Recognition of Daily Life Pattern based on Event Calculus

배석현(Seok-Hyun Bae),방성혁(Sung-hyuk Bang),박현규(Hyun-Kyu Park),전명중(Myung-Joong Jeon),김제민(Je-Min Kim),박영택(Young-Tack Park) Korean Institute of Information Scientists and Eng 2018 정보과학회논문지 Vol.45 No.5

Current Status of Collaboration and Research Trend in Microbiology in South Korea

Seok Hyun Bae(배석현),Keuntae Cho(조근태) 한국경영과학회 2018 經營 科學 Vol.35 No.4

The objective of the study is to identify how researchers conduct collaborative work and which is the trend in the field of Korean microbiology using scientometrics and social network analysis (SNA) approaches from Korea-based microbiology articles. For the purpose, we use co-authorship as an indicator of the research collaboration network by analyzing 7,756 published sources from representative Korean microbial articles between 1996 and 2015. The results show that cooperation with different institutions is gradually increasing over time but the proportion of co-authorship within an institution decreased. And the average number of authors per paper is 4.928 during the last two decades in the Korean Microbiology Journal. Moreover, microbiology research trends in Korean have shifted with the passage of time from microbes and enzymes themselves to genomes, recombinant methods, proteins, bio-engineering, immunity, and molecular diagnosis. The significance of this study is that it is the first attempt to analyze the trends in Korean microbiology research using scientometrics with representative microbiology articles published by Korean scholars and using co-occurrent keywords to identify the research trend of Korean microbiology.

순환신경망 기반의 사용자 의도 예측 모델

방성혁,배석현,박현규,전명중,김제민,박영택 한국정보과학회 2018 정보과학회논문지 Vol.45 No.4

Several studies have been conducted on human intention prediction with the help of machine learning models. However, these studies have indicated a fundamental shortcoming of machine learning models since they are unable to reflect a long span of past information. To overcome this limitation, this paper proposes a human intention prediction model based on a recurrent neural network(RNN). For performing predictions, the RNN model classifies the patterns of time-series data by reflecting previous sequence patterns of the time-series data. For performing intention prediction using the proposed model, an RNN model was trained to classify predefined intentions by using attributes such as time, location, activity and detected objects in a house. Each RNN node is composed of a long short-term memory cell to solve the long term dependency problem. To evaluate the proposed intention prediction model, a data generator based on the weighted-graph structure has been developed for generating data on a daily basis. By incorporating 23,000 data instances for training and testing the proposed intention prediction model, a prediction accuracy value of 90.52% was achieved. 기계 학습 모델 구축을 통한 인간의 의도 예측은 기존에도 제공되어 왔으나, 특정 행위가 발생하는 시점으로부터 먼 과거의 정보를 반영한 의도 예측이 어렵다는 단점이 존재했다. 이 문제점의 극복을 위해, 본 논문에서는 순환 신경망(RNN – Recurrent Neural Network) 기반의 행위 의도 예측 모델 학습 기법을 제안한다. 순환 신경망 모델은 시계열(Time-Series) 데이터의 패턴을 분석하여 과거의 시점이 반영된 예측 결과를 생성한다. 본 논문이 제안하는 순환 신경망 기반의 의도 예측 모델은 시간, 공간, 행위, 물체, 의도로 구성된 생활 데이터 시퀀스를 바탕으로 사용자의 의도를 예측할 수 있도록 학습된다. 순환 신경망의 각 노드는 의도 예측 모델이 먼 과거의 데이터 시퀀스를 고려하여 의도를 예측 할 수 있도록 LSTM(Long-Short Term Memory) Cell로 구성하였다. 순환 신경망 기반의 의도 예측 모델의 성능 평가를 위해, 본 논문에서는 행위 의도에 대한 가중치 그래프 기반 데이터 생성기를 구축하여 실제 실내에서 발생하는 인간 활동에 가까운 데이터를 자동으로 생성하여 실험에 사용했다. 총 23,000개의 데이터가 의도 모델 학습과 검증에 사용되었으며, 학습된 모델의 의도 예측 정확도 측정 실험을 한 결과로 평균 90.52%의 예측 정확도를 보였다.