영한 기계번역을 위한 구구조 규칙의 수립 : 병렬 복합어 어순을 중심으로 Concentrating on the Ordering of Frozen Pairs

박희문 한국현대영어영문학회 1990 현대영어영문학 Vol.32 No.1

The purpose of the study is to present Linear Precedence statements of English and Korean "frozen pairs" in the sense of GPSG, which is vital to English-Korean machine translation. The ordering of frozen pairs of both languages was contrasted with each other on the basis of semantic correspondence, whose results are expected to reveal important clues of human cognitive systems as well as to contribute to seeking linguistic universals. One major problem in the course of contrastive analysis works is how to handle the lexical disparates between English and Korean, due to different social and cultural backgrounds. To make it more objective, the study mainly deals with semantically equivalent "frozen, pairs" from both languages, ineluding the ones whose semantic categories are similar each to each in a broader sense. The study also deals with the vocabularies with Chinese etymology which are now being used like those of Korean originals. The study results can be summarized as follows: 1. Generally speaking, the ordering of frozen pairs of both languages seems to conform to "Me-first Principle" by Cooper & Ross, arguing that Me is a male adult human who is located in space and time at the speech event. 2. The semantic factors which show prominent ordering contrasts between the two languages belong to the following categories : Time vs. Space, Right vs. Left, Upper vs. Lower clothes, and Vertical vs. Horizontal. These contrasts seem to suggest that people process those dimensions differently from culture to culture, which is known to be as a "marked" phenomenon. 3. The study reveals that, in both languages, positive elements precede negative elements in common, and their opposite odering can often be found when they are used within an emphasized context. 4. The study also shows that people tend to regard the following semantic categories as "unmarked" ; namely, Friendly, Up, Front, Internal, Singular, Solid, Large, Long, and Strong, upon which human value systems and thinking processes seem to be based. 5. The simply juxtaposed expressions with unknown semantic causes are possible to be analyzed through phonological rules by Ross(1982). It seems, however, that semantic rules tend, to override phonological rules, and the latter seems to be more appropiate for explaining the ordering of twin words. Further studies are needed particulary on how to determine the semantic categories of frozen paires in details in order for this sort of study to be carried out more effectively.

Studies on Adsorption of Organic Colorants to Microporous materials

박희문,박수열,전근,오세화 한국공업화학회 2000 한국공업화학회 연구논문 초록집 Vol.2000 No.0

N/A N/A

신경회로망을 사용한 물고기 로봇의 빠른 방향 전환 궤적 설계

박희문,박진현,Park, Hee-Moon,Park, Jin-Hyun 한국정보통신학회 2014 한국정보통신학회논문지 Vol.18 No.3

본 연구에서는 유체 속에서의 로봇의 방향전환 메커니즘의 성능을 개선하고 최적화하기 위하여 물 속 자연환경에 최적화되어 있는 물고기의 CST(CST:C-shape sharp turn) 패턴을 모방하여 물고기 로봇의 꼬리 관절 궤적을 신경회로망(neural network)을 사용하여 제안하였다. 물고기의 CST 패턴을 모방하기 위해 CST 패턴을 순차적으로 기록한 정보를 수치적으로 변환하여 좌표 데이터를 생성하고 함수화하였다. 함수화된 모션 함수를 물고기 로봇의 상대 관절각으로 변환하였으나, 구해진 상대 관절 궤적은 잉어의 순차적 기록에 의해 구해진 각도이므로 분해능이 떨어져 실제 물고기 로봇의 제어에 적용하기 어렵다. 그러므로 상대 관절 궤적을 일반화 기능이 뛰어난 신경회로망을 사용하여 보간하고 물고기 로봇에 적용하였다. 모의실험을 통하여 신경회로망을 이용한 상대 관절 궤적 함수가 고차의 다항식 궤적 함수에 비하여 물고기 로봇의 CST 모션에 더 좋은 성능을 나타냄을 확인하였다. In this study, in order to improve and optimize the performance of the turning mechanism for a fish robot in the fluid, we propose the tail joint trajectories using neural networks to mimic the CST(C-shape Sharp Turn) patterns of a real fish which is optimized in the natural environment. In order to mimic the CST patterns of a fish, we convert the sequential recording CST patterns into the coordinate data, and change the numerical coordinate data into a functions. We change the motion functions to the relative joint angles which is adapted to suit robot's shape and data. However, these relative joint trajectories obtained by the sequential recording of the carp have low-precision. It is difficult to apply to the control of a fish robot. Therefore, the relative joint trajectories are interpolated using neural networks with superior generalization ability and applied to the fish robot. we have found that the proposed method using neural networks is superior to ones using high-order polynomial equation through the computer simulations.

Triton X-100이 Trichoderma koningii의 성장 및 형태에 미치는 영향

박희문,민경렴,맹필재,하영칠 한국미생물학회 1991 미생물학회지 Vol.29 No.5

We investigated the effect of Triton X-100 on the growth and morphology of Trichoderma koningii by comparing various parameters representing the frowth of mold in the presence or absence of Triton X-100. The specific growth rate and doubling time of T. koningii were not affected by the addition of 0.05% Triton X-100 in batch culture. However, in the presence of Triton X-100, cultures reached its stationary phase earlier and showed reduced level in total yield of biomass. The addition of Triton X-100 into solid medium also resulted in decrease in the colony radial growth rate and this response was correlated with the formation of mycelia which showed increase in branching and septation in the presence of Triton X-100.

F-29 Prevalence of Airway Disease in Patients with Idiopathic Pulmonary Fibrosis

박희문,최선미 대한결핵 및 호흡기학회 2017 대한결핵 및 호흡기학회 추계학술대회 초록집 Vol.124 No.-

Background: Idiopathic pulmonary fibrosis (IPF) is a progressive fibrosing interstitial pneumonia whose main symptoms are dyspnea and cough. Asthma, eosinophilic bronchitis (EB), and chronic obstructive pulmonary disease (COPD) are airway diseases which provoke dyspnea and cough. Because of no study of the prevalence of asthma and EB in IPF, we performed this study to evaluate that of airway disease including asthma, EB, and COPD in IPF. Method: This study is a single-institution prospective study, from June 2017 to September 2017 at the Seoul national university hospital. Spirometry with bronchodilator, metacholine provocation test, Induced sputum with eosinophil stain, multiple antigen simultaneous test (MAST), exhaled nitric oxide (FENO), serum IgE, eosinophil count, were performed to evaluate the presence of airway disease. Result: 26 patients were enrolled in this study. The prevalence of airway disease was 34.6% (9/26). 7 patients were diagnosed with EB (26.9%), 2 with COPD (7.7%), and none with asthma. Patients with predisposition to allergy were 13 (50%) among whom, 11 are positive for MAST (42.3%), 2 with high FENO (7.7%), 6 with increasing IgE (23.1%), one with eosinophilia (3.9%). Questionnaire for quality of life (CAT, SGRQ) and symptom severity (mMRC, 6-minute walk test, EQ-5D-VAS, CQLQ) were not different regardless of combined airway disease. Conclusion: Our study suggests that EB and COPD are prevalent in IPF, while asthma is not. It is necessary to conduct research on a larger number of patients in the future.

고내구성 교면포장 아스팔트 혼합물의 공용성 평가를 위한 포장가속시험

박희문,최지영,이현종,황의윤,Park, Hui-Mun,Choe, Ji-Yeong,Lee, Hyeon-Jong,Hwang, Ui-Yun 한국도로학회 2008 도로 : 도로학회지 Vol.10 No.1

컨벌루션 신경망을 사용한 다중 차선 인식

박희문,황광복,배준경,박진현 한국기계기술학회 2022 한국기계기술학회지 Vol.24 No.2

In this study, the multi-lane detection problem is expressed as a CNN-based regression problem, and the lane boundary coordinates are selected as outputs. In addition, we described lanes as fifth-order polynomials and distinguished the ego lane and the side lanes so that we could make the prediction lanes accurately. By eliminating the network branch arrangement and the lane boundary coordinate vector outside the image proposed by Chougule’s method, it was possible to eradicate meaningless data learning in CNN and increase the fast training and performance speed. And we confirmed that the average prediction error was small in the performance evaluation even though the proposed method compared with Chougule’s method under harsher conditions. In addition, even in a specific image with many errors, the predicted lanes did not deviate significantly, meaningful results were derived, and we confirmed robust performance.

단일 원형 경계상자를 사용한 YOLO 기반 궐련 담배 검출 시스템

박희문,전향식,황광복,박진현 한국정보통신학회 2023 한국정보통신학회논문지 Vol.27 No.8

In modern factory automation, various cutting-edge technologies such as mechanical, electrical, electronic, and computer engineering are applied in the manufacturing process to increase production efficiency and improve product quality. In this study, a YOLO-based cigarette detection system using a single circular bounding box is proposed to accurately detect defective packaging, such as missing cigarettes or reverse alignment in the cigarette production process. In particular, Proposed-Net2 shows an average precision similar to that of the YOLOv4-Tiny network in terms of performance, uses 10% of memory compared to the existing ones, and detects cigarettes with about twice as fast processing time. And it was confirmed that effective learning is possible by augmenting various types of data with only a small number of training images by combining object-based cropped images with rotation and randomly generated bounding boxes for insufficient training data. This is applicable even when hardware specifications are low, such as an embedded environment. 현대의 공장자동화 분야에서는 생산효율 증대와 생산 제품의 품질 향상을 위해 제조 공정에서 기계, 전기, 전자, 컴퓨터 공학 등의 다양한 최신 기술이 적용되고 있다. 본 연구에서는 궐련 담배 생산 공정에서 궐련 담배의 누락이나 역방향 정렬과 같은 불량 포장을 정확하게 검출하기 위해 단일 원형 경계상자를 사용한 YOLO 기반 궐련 담배 검출 시스템을 제안하였다. 특히, Proposed-Net2는 성능 면에서 YOLOv4-Tiny 네트워크와 유사한 평균 정밀도를 나타내며, 기존 대비 10% 수준의 메모리를 사용하며 약 2배 정도의 빠른 처리 시간으로 궐련 담배를 검출하였다. 그리고 부족한 학습 데이터는 객체 기반으로 잘라낸 이미지를 회전 및 무작위로 생성한 경계상자에 합성하여 적은 수의 학습 이미지만으로 다양한 형태의 데이터를 증강하여 효과적인 학습이 가능함을 확인하였다. 이는 임베디드 환경과 같은 하드웨어 사양이 낮은 곳에서도 적용할 수 있는 장점이 있다.

도로교통 분야에서의 자산관리

박희문,Park, Hui-Mun 한국도로학회 2004 도로포장공학회지 Vol.6 No.1

다단계 FWD 하중을 이용한 블록포장의 비선형 거동 분석

박희문,김연태,이수형 한국도로학회 2016 한국도로학회논문집 Vol.18 No.6

OBJECTIVES : The objective of this study is to analyze the nonlinear behavior of block pavements using multi-load level falling weight deflectometer (FWD) deflections. METHODS: Recently, block pavements are employed not only in sidewalks, but also in roadways. For the application of block pavements in roadways, the structural capacities of subbase and subgrade are important factors that support the carry traffic load. Multi-load level FWD testing was conducted on block pavements to analyze their nonlinear behavior. The deflection ratio due to the increase in load was analyzed to estimate the nonlinearity of block pavements. Finite element method with nonlinear soil model was applied to simulate the actual nonlinear behavior of the block pavement under different levels of load. RESULTS: The results of the FWD testing show that the center deflections in block pavements are approximately ten times greater than that in asphalt pavements. The deflection ratios of the block pavement due to the increase in the load range from 1.2 to 1.5, indicating that the deflection increased by 20~50%. The material coefficients of the nonlinear soil model were determined by comparing the measured deflections with the predicted deflections using the finite element method. CONCLUSIONS: In this study, the nonlinear behavior of block pavements was reviewed using multi-load level FWD testing. The deflection ratio proposed in this study can estimate the nonlinearity of block pavements. The use of nonlinear soil model in subbase and subgrade increases the accuracy of predicting deflections in finite element method.