http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
영상기반 주차공간 분류 딥 모델을 위한 데이터 증강기법
유호진(Hojin Yoo),전경구(Kyungkoo Jun) Korean Institute of Information Scientists and Eng 2022 정보과학회논문지 Vol.49 No.2
A parking occupancy state determination system using an ultrasonic sensor or a camera is mainly used in indoor parking lots. However, in the case of an outdoor parking lot, there is a limit to the introduction of these systems due to the high installation cost and accuracy problems. In addition, the application of deep learning is restricted because it is difficult to obtain representative learning data due to diverse lighting conditions, camera positions, and features. In this paper, we analyzed the effect of augmentation techniques on the performance of a deep model for parking status classification in such a data shortage situation. To this end, the parking area images were classified by situations. Four augmentation techniques were applied to the training of ResNet, EfficientNet, and MobileNet. Based on performance evaluation, the accuracy was improved by up to 5.2%, 8.67%, and 15.44%p in the case of mixup, stopper, and rescaling methods, respectively. On the other hand, in the case of center crop, which was known to have performance improvement in other studies, the accuracy decreased by an average of 4.86%p.
통합마스트-송수신 모듈의 열화인자 분석을 통한 결함진단 및 잔여수명 예측 연구
고호진(Hojin Koh),이종학(JongHak Lee),라영은(YoungEun Ra),강민성(Minsung Kang),유영민(Yeongmin Yoo),권대일(Daeil Kwon),이종수(Jongsoo Lee) (사)한국CDE학회 2020 한국CDE학회 논문집 Vol.25 No.2
Frequently installed on latest destroyers, integrated mast (I-mast) conducts tasks including coast and air-space surveillances. I-mast is a complex multi-domain system with various mechanical and electrical subsystems and components. In the perspective of diagnosis and maintenance, it is important to technologically manage this system for long operating time periods and eventually increase its survivability. We explore the prognostics and health management (PHM) process of I-mast and transmitter receiver module (TRM) using the framework of Modelica-based system simulation. With Modelica, one can reuse components employed in a model, easily model a system expressed in mathematical language, and conduct a multi-dimensional, multi-physics analysis. First, failure mode effects and criticality analysis (FMECA) was conducted to extract the major failure and degradation factors of an I-mast structure and TRM under the vessel’s vibrating and operating conditions. To quantitatively assess the risk priority number (RPN), we propose the use of design of experiments (DoE) and analysis of variance (ANOVA). To preprocess the vibration acceleration signals, principal component analysis (PCA) was used to identify major features from statistical multi-components. Subsequently, based on the diagnosis of fault states, the prediction of remaining useful life (RUL) of I-mast and TRM has been discussed in the context of PHM.
Compensation method for respiratory motion in proton treatment planning for mobile liver cancer
Jeong, Hojin,Lee, Se Byeong,Yoo, Seung Hoon,Lim, Young Kyung,Kim, Tae Hyun,Park, Seyjoon,Chai, Gyu Young,Kang, Ki Mun,Shin, Dongho John Wiley and Sons Inc. 2013 Journal of applied clinical medical physics Vol.14 No.2
<P>We evaluated the dosimetric effect of a respiration motion, and sought an effective planning strategy to compensate the motion using four‐dimensional computed tomography (4D CT) dataset of seven selected liver patients. For each patient, we constructed four different proton plans based on: (1) average (AVG) CT, (2) maximum‐intensity projection (MIP) CT, (3) AVG CT with density override of tumor volume (OVR), and (4) AVG CT with field‐specific proton margin which was determined by the range difference between AVG and MIP plans (mAVG). The overall effectiveness of each planning strategy was evaluated by calculating the cumulative dose distribution over an entire breathing cycle. We observed clear differences between AV G and MIP CT‐based plans, with significant underdosages at expiratory and inspiratory phases, respectively. Only the mAVG planning strategy was fully successful as the field‐specific proton margin applied in the planning strategy complemented both the limitations of AVG and MIP CT‐based strategies. These results demonstrated that respiration motion induced significant changes in dose distribution of 3D proton plans for mobile liver cancer and the changes can be effectively compensated by applying field‐specific proton margin to each proton field.</P><P>PACS numbers: 87.55.D; 87.53.Bn; 87.53.Jw; 87.55.dk</P>
모바일 앱 프로파일 기반 센서 레지스트리 시스템의 구현 및 평가
최호진 ( Hojin Choi ),유현석 ( Hyunseok Yoo ),정동원 ( Dongwon Jeong ),전근환 ( Keunhwan Jeon ) 한국정보처리학회 2016 한국정보처리학회 학술대회논문집 Vol.23 No.1
이 논문에서는 기존 센서 레지스트리 시스템의 처리 성능을 개선하기 위해 제안된 모바일 앱 프로파일 기반 센서 필터링 기법을 실험하고 평가한다. 이기종 센서 네트워크 환경에서 센서 레지스트리 시스템은 센서 메타데이터를 제공함으로써 센서 정보의 의미 처리를 즉시적으로 가능하게 한다. 하지만 불필요한 센서 메타데이터까지 처리하여 전체적인 처리 속도를 저하시킨다. 이 문제점을 개선하기 위해 모바일 앱 프로파일 기반 센서 필터링 기법을 제안한다. 모바일 앱 프로파일을 활용하여 센서 데이터의 유효성을 식별 후 모바일 기기에 센서 메타데이터를 기록한다. 기록된 정보를 재사용하여 불필요한 센서 데이터를 필터링 한다. 불필요한 센서 데이터를 처리 하지 않고, 센서 메타 데이터의 요청 횟수를 줄여 전체적인 센서 데이터 처리 속도를 향상 시킨다. 기존 방법과 제안 방법을 구현 하고 실험하여 제안 방법의 전체적인 처리 속도가 향상됨을 확인한다.
Noise Covariance Adaptation of Doppler Velocity Log Considering Vehicle’s Attitude
Jaehyuck Cha,Hojin Ju,Chan Gook Park,Kijeong Yoo,Chanju Park 제어로봇시스템학회 2018 제어로봇시스템학회 국제학술대회 논문집 Vol.2018 No.10
Doppler velocity Log (DVL) is a sensor offering velocity information. This paper analyzes the error characteristics of DVL measurements and proposes a noise covariance adaptation method by using the vehicle’s attitude information. For the successful operation of various marine vehicles, proper navigation systems are required. Inertial navigation system (INS) is usually adopted as a main navigation system in many applications, but it has a characteristic of cumulatively increasing error with time. However, the errors can be suppressed by using another sensor such as DVL. The accuracy of DVL measurements is dependent to the distances to the sea floor along the transducer axes, and each distance is a function of the vehicle’s attitude. The analysis is verified by computer simulations, and the result shows that the proposed method decreases the navigation errors.
Development of a TPU/CNT/Cu Composite Conductive Filament with a High CNT Concentration
Hansol Kim,Hojin Yoo,Seungcheol Shin,Jungho Cho,Sang-Woo Han,In Hwan Lee 한국정밀공학회 2023 International Journal of Precision Engineering and Vol.24 No.2
In this study, a flexible conductive filament was fabricated by mixing thermoplastic polyurethane (TPU), carbon nanotubes (CNTs), and Cu powder. A polymer extruder was used to disperse conductive materials in the TPU matrix. Because the dispersion of CNTs in polymers is difficult, the dispersion process was repeated several times for the homogeneity of the fabricated conductive filament. Cu powder with high electrical conductivity was additionally added to improve the electrical characteristics of the conductive polymer. As TPU generally has excellent ductility and durability, the fabricated filament can secure flexibility. The resistance was measured and compared according to the mixing ratio of CNT/Cu powder. Based on the obtained results, the resistance of the conductive filament decreased as the CNT and Cu powder contents were increased. The addition of metal powders, such as CNTs and Cu powder, however, reduced the ductility of TPU. Finally, the developed conductive filament was used to fabricate a simple closed photodiode circuit.