제약적 최소자승추정을 이용한 미분기반 다중계층 광류 추정

김영춘,신용달,권기구 永同大學校 2001 硏究論叢 Vol.7 No.1

본 논문에서는 광류 제약 방정식(optical flow constraint equation) 및 영상의 공간적인 상관성을 고려하여 제약적 최소자승추정(constrained least square estimation)을 이용한 미분기반 다중계층(multiscale) 광류 추정 방법을 제안하였다. 제안한 방법은 인접한 화소간의 움직임 속도가 유사하다는 가정 하에 다중계층으로 광류를 추정하였다. 또한 본 논문에서는 광류 제약 방정식 및 인접한 화소간의 움직임 속도가 유사하다는 가정에 의하여 발생하는 오차를 줄이기 위한 정제(refinement) 방법도 제안하였다. 제안한 방법이 기존의 방법에 비하여 더욱 정확하게 광류를 추정할 수 있음을 실험을 통하여 확인하였다. In this paper, we propose a gradient-based multiscale optical flow estimation method using the constrained least-square estimation, considering the spatial correlation of the image. The basic assumption of the proposed method is that the neighboring pixels are having a similar velocity. Also we propose a refinement method to decrease the error, reviving optical flow constraint equation and neighboring pixels are having a similar velocity. Experimental results show that the proposed method can more accurately estimate the optical flow than the conventional methods.

노화가 인체 중간엽 줄기세포로부터 조골세포로의 증식 및 분화에 미치는 영향

백기현,태현정,오기원,이원영,조정기,권순용,강무일,차봉연,이광우,손호영,강성구,김춘추 대한내분비학회 2003 Endocrinology and metabolism Vol.18 No.3

연구배경: 일반적으로 골다공증과 연관된 위험인자로는 연령, 폐경, 약물, 불충분한 칼슘섭취, 만성질환 및 운동부족 등이 있는데, 특히 노화가 진행할수록 골밀도가 감소하는 것은 잘 알려져 있다. 노화와 관련하여 진행되는 골소실은 조골세포 및 전구조골세포의 기능적 결핍에 의한 골형성의 감소가 주요한 요인으로 여겨지고 있다. 그 동안 연령이 조골모 세포의 양과 조골모 세포로부터 성숙조골세포로의 분화 및 증식에 미치는 영향에 대한 일부 보고들이 있었으나 아직 일치된 견해는 없는 형편이다. 방법: 다양한 연령의 사람으로부터 골수를 채취, 중간엽 줄기세포가 포함된 단핵세포를 분리한 후 조골세포로 분화하기 좋은 조건하에서 배양하였다. 대상군은다시 젊은군과 노령군으로 구분하여 다양한 변수를 비교 분석하였다. 일차배양에서는 CFU-F를 계수하여 골수내 중간엽 줄기세포의 수를 추산하였고, 칼슘측정을 통하여 기질의 무기화 정도를 비교하였다. 계대배양후 이차배양에서는 시기별로 알카리성 포스파타제 활성도를 측정하고 오스테오칼신 mRNA의 발현을 관찰하여 젊은군과 노령군 사이의 증식능 차이를 비교하였다. 또한 이차배양 시기별로 MTT 측정을 하여 양군간에 증식능 차이가 있는지 알아보았다. 결과: 1. 일차배양 15일째에 평균 CFU-F의 수는 젊은군에서 유의하게 많았다(젊은군 148.3±28.9, 노령군 54.3±9.1, p=0.02). CFU-F의 평균면적은 젊은군에서 넓은 경향을 보였으나 통계적으로 유의하지는 않았다. 2. 일차배양 17일 경과 후 양군간에 기질 칼슘 침착정도는 유의한 차이를 보이지 않았다(젊은군 103.6±50.6, 노령군: 114.0±56.5, p=NS). 3. 이차배양 10일째에 젊은군에서 알카리성 포스파타제 활성도가 고령군에 비해 유의하게 높았다(젊은군: 935.5±115.0 U/mg, 노령군: 578.4±115.7U/mg,p.0.05). 고령군에서는 시간 경과에 따른 변화가 미약했으며 전반적으로 알카리성 포스파타제의 활성도가 젊은군에 비해 낮았다. 4. 이차배양도중 오스테오칼신 mRNA의 발현은 배양시기별로 젊은군에 비해 고령군에서 더 낮은 경향을 관찰할 수 있었다. 5. 이차배양 10일과 15일에 젊은군에서 노령군보다세포증식이 유의하게 증가된 양상을 보였다(10앓 젊은군 0.73±0.05, 노령군 0.58±0.04, p=0.05, 15일; 젊은군 0.80±0.05, 노령군 0.70±0.03, p=0.05).결론: 이상의 연구에서 저자들은 노령군에서 젊은군보다 골수 내 중간엽줄기세포의 수가 적고, 노령군에서 유래한 전구조골세포의 성숙조골세포로의 증식 및 분화가 젊은군 보다 감소해 있는 것을 관찰할 수 있었다. Background: Osteoblasts originate from osteoprogenitor cells in bone marrow stroma, termed mesenchymal stem cells (MSCs) or bone marrow stromal cells. Each MSC forms colonies (colony forming units-fibroblasts [CFL-Fs]) when cultured ex vivo. There are some reports about the age-related changes of the number and osteogenic potential of osteoprogenitor cells, but any relationship has not been clearly established in humans. In this study, we counted MSCs using CFU-Fs count and examined the proliferative capacity and differentiation potential of osteoprogenitor cells. Finally, we analyzed how these parameters varied with donor age. Methods: Bone marrow was obtained from the iliac crest of young (n=6, 27.2±8.6 years old) and old (n= 10, 57.4k6.7 years old) healthy donors. Mononuclear cells, including MSCs, were isolated and cultured in osteogenic medium. In primary culture, we compared the colony-forming efficiency of MSCs between the two groups and determined the matrix calcification. When primary culture showed near confluence, the cells were subcultured. Alkaline phosphatase activity, osteocalcin expression by RT-PCR and proliferative potential by MTT assay were examined by the time course of secondary culture. Results: At the 15th day of primary culture, the mean number of CFU-Fs was significantly higher in the younger donors (young: 148.3±28.9, old: 54.3±9.1, p=0.02) and the mean size of CFL-Fs was also larger in the younger donors than the older donors. However, matrix calcification was not different between the two groups (young: 103.6±50.6, old: 114.0±56.5, p=NS). In secondary culture, alkaline phosphatase activities were significantly lower in the older donors. The younger donors showed peak alkaline phosphatase activity at day 10, while the older donors didn't showed a remarkable peak (young: 935.5±115.OU/mg, old: 578.4±115.7U/mg, p<0.05). Total cell number as a proliferative index increased progressively during the secondary culture and a significantly greater cell number was noted in the younger donors. Osteocalcin expression was generally upregulated in the younger donors, but this was not statistically significant. Conclusion: Our study shows that the number of osteoprogenitor cells is decreased during aging and that the proliferative capacity and differentiation potential of osteoprogenitor cells seem to be reduced during aging (J Kor SOC Endocrinol 18:296-305, 2003).

HMM-Based Transient Identification in Dynamic Process

Kee-Choon Kwon 제어로봇시스템학회 2000 Transaction on control, automation and systems eng Vol.2 No.1

TECHNICAL REVIEW ON THE LOCALIZED DIGITAL INSTRUMENTATION AND CONTROL SYSTEMS

Kwon, Kee-Choon,Lee, Myeong-Soo Korean Nuclear Society 2009 Nuclear Engineering and Technology Vol.41 No.4

This paper is a technical review of the research and development results of the Korea Nuclear Instrumentation and Control System (KNICS) project and Nu-Tech 2012 program. In these projects man-machine interface system architecture, two digital platforms, and several control and protection systems were developed. One platform is a Programmable Logic Controller (PLC) for a digital safety system and another platform is a Distributed Control System (DCS) for a non-safety control system. With the safety-grade platform PLC, a reactor protection system, an engineered safety feature-component control system, and reactor core protection system were developed. A power control system was developed based on the DCS. A logic alarm cause tracking system was developed as a man-machine interface for APR1400. Also, Integrated Performance Validation Facility (IPVF) was developed for the evaluation of the function and performance of developed I&C systems. The safety-grade platform PLC and the digital safety system obtained approval for the topical report from the Korean regulatory body in February of 2009. A utility and vendor company will determine the suitability of the KNICS and Nu- Tech 2012 products to apply them to the planned nuclear power plants.

HMM-Based Transient Identification in Dynamic Process

Kwon, Kee-Choon Institute of Control 2000 Transaction on control, automation and systems eng Vol.2 No.1

In this paper, a transient identification based on a Hidden Markov Model (HMM) has been suggested and evaluated experimentally for the classification of transients in the dynamic process. The transient can be identified by its unique time dependent patterns related to the principal variables. The HMM, a double stochastic process, can be applied to transient identification which is a spatial and temporal classification problem under a statistical pattern recognition framework. The HMM is created for each transient from a set of training data by the maximum-likelihood estimation method. The transient identification is determined by calculating which model has the highest probability for the given test data. Several experimental tests have been performed with normalization methods, clustering algorithms, and a number of states in HMM. Several experimental tests have been performed including superimposing random noise, adding systematic error, and untrained transients. The proposed real-time transient identification system has many advantages, however, there are still a lot of problems that should be solved to apply to a real dynamic process. Further efforts are being made to improve the system performance and robustness to demonstrate reliability and accuracy to the required level.

Accident Identification in Nuclear Power Plants using AI Technology

Kee-Choon Kwon,Dong-Young Lee 제어로봇시스템학회 1999 제어로봇시스템학회 각 지부별 자료집 Vol.1999 No.12

원전 디지털 원자로보호계통 소프트웨어 안전보증 패러다임 적용 및 분석

권기춘(Kee-Choon Kwon),이장수(Jang-Soo Lee),지은경(Eunkyoung Jee) 한국정보과학회 2017 정보과학회 컴퓨팅의 실제 논문지 Vol.23 No.6

원자력발전소 안전-필수 소프트웨어를 개발하고 검증 및 확인을 수행하여 규제기관으로부터 인허가를 받기 위하여 단순하게 문서를 읽고 검토해서는 개발, 구현 및 검증활동에 대한 신뢰성과 안전성 확보에 대하여 정확하게 판단하기가 쉽지 않다. 따라서 이러한 활동, 특히 안전보증 활동이 소프트웨어 결함이 허용 가능한 수준인지 판단하기 위한 체계적인 평가기술이 필요하다. 본 연구에서는 원전 디지털 원자로보호계통의 비교논리 프로세서와 동시논리 프로세서를 대상으로 제작자가 수행한 개발 및 검증 결과물의 수준과 깊이를 평가하기 위해 안전진술(Safety case) 방법론을 적용하고 그 결과를 분석한다. 안전진술 방법론 적용으로 기존의 안전입증 방법을 효과적으로 보완할 수 있음을 확인하였다. In the verification and validation procedures regarding the safety-critical software of nuclear power plants for the attainment of the requisite license from the regulatory body, it is difficult to judge the safety and dependability of the development, implementation, and validation activities through a simple reading and review of the documentation. Therefore, these activities, especially safety assurance activities, require systematic evaluation techniques to determine that software faults are acceptable level. In this study, a safety case methodology is applied in an assessment of the level and depth of the results of the development and validation of a manufacturer in its targeting of the bistable processor of a digital reactor protection system, and the evaluation results are analyzed. This study confirms the possibility of an effective supplementation of the existing safety demonstration method through the application of the employed safety case methodology.

원전 디지털 원자로보호계통 소프트웨어 안전보증 패러다임

권기춘(Kee-Choon Kwon),이장수(Jang-Soo Lee),박기용(Gee-Yong Park),지은경(Eunkyoung Jee) 한국정보과학회 2016 한국정보과학회 학술발표논문집 Vol.2016 No.6

시용 유기물의 종류가 고추 재배지 토양 미생물상에 미치는 영향

박기춘(Kee-Choon Park),김영숙(Yeong-Suk Kim),권오훈(Oh-Hoon Kwon),권태룡(Tae-Ryong Kwon),박상구(Sang-Gu Park) 한국토양비료학회 2008 한국토양비료학회지 Vol.41 No.2

Diverse organic amendments available in local areas have been used to improve soil quality in red pepper field and so the need for investigating the soil chemical and biological properties changed by the organic amendments application is increasing. Soil microbial diversities were measured by phospholipid fatty acid (PLFA) and Biolog EcoPlateTM. Compost was most effective for improving soil chemical properties including pH, EC, total nitrogen, P, K, and Ca, and bark increased soil organic matter significantly (P=0.05). Compost increased the fatty acids indicating actinomycetes and vascular arbuscular fungi, and ratio of cy19:0/18:1w7c and monounsaturated fatty acids/saturated fatty acids in soils in PLFA analysis. Bark increased soil fungal indicators in PLFA analysis (P=0.05). Principal component analysis of Biolog EcoPlate data and PLFA differentiated the compost- and bark-amended soils from other organic matteramended soils especially the soil incorporated with compost. More researches are needed to use bark for improving soil microbial properties because the soil chemical and microbiological properties caused by compost and bark are significantly different.

은닉 마르코프 모델을 이용한 원자력발전소 사고진단

권기춘(Kee-Choon Kwon),김진형(Jin-Hyung Kim) 한국정보과학회 1997 정보과학회논문지(B) Vol.24 No.12

원자력발전소와 같이 복잡한 시스템에서 사고상태를 조기에 진단하여 적절한 운전조치를 취하는 것이 필요하다. 원자력발전소에서 사고가 발생하면 주요변수 및 기기의 증상패턴에 따라 사고유형을 패턴으로 표시할 수 있으며, 이 패턴을 인식함으로써 사고의 종류를 진단할 수 있다. 따라서 시간적인 특성과 공간적인 특성을 효과적으로 모델링할 수 있는 이중 확률 모델인 HMM (Hidden Markov Model)을 적용할 수 있게 된다. 학습용 데이터는 시험용 시뮬레이터로부터 획득하여 자기조직화지도를 이용하여 벡터양자화된다. 각 사고에 대해서 학습에 의하여 하나의 모델이 생성되며, 학습은 최대확률 파라미터 추정 방법을 사용하는데, Forward-Backward 알고리즘과 Baum-Welch 재추정 알고리즘을 이용한다. 진단은 주어진 입력패턴에 내해서 가장 큰 확률 값을 가지는 모델로 결정하며, Viterbi 알고리즘으로 각 모델에 대해 최적의 경로를 구한 다음, 최적 경로에 따른 확률을 구한다. HMM을 이용한 사고진단시스템은 left-to-right 모델을 사용하였으며, 8개의 사고와 정상상태를 진단하기 위하여 22개의 입력변수를 사용하였다. 모의 실험에서 8개의 사고유형 모두 정확하게 진단하였으며,입력패턴이 센서의 오류, 일부 기기의 오동작등으로 하나의 변수값이 부정확한 경우에도 거의 정확하게 진단하였다. Classification of the types of accidents at an early stage of an accident in nuclear power plants is crucial for proper action selection. A plant accident can be classified by its symptom patterns related to the principal variables and operating status of major equipment. The Hidden Markov Model (HMM), a double stochastic process enables modeling of, not only spatial phenomena, but also time scale distances. The training data from a test nuclear simulator are vector quantized by the Self-Organizing Map, an unsupervised artificial neural network clustering algorithm. The HMM is created for each accident from a set of training data which are the result of vector quantization The HMM is trained by the maximum-likelihood estimation method which uses a forward-backward algorithm and a Baum-Welch re-estimation algorithm. The accident diagnosis is decided by calculating which model has the highest probability for given test data. The optimal path for each model at the given observation is found by the Viterbi algorithm, the probability of optimal path is then calculated. The system uses a left-to-right model including 6 states and 22 input variables to classify 8 types of accidents and a normal state. The simulation results show that the proposed system identifies the accident types correctly. It is also shown that the diagnosis is performed well for incomplete input observation caused by sensor fault or malfunction of certain equipment.