http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
이용빈(Lee, Yongbin),이주영(Lee, Juyoung),인진환(In, Jinhwan),이지은(Lee, Jeeun),조용훈(Cho, Yonghoon) 한국주거학회 2020 한국주거학회 학술대회논문집 Vol.32 No.2
The purpose of the research is to discover the social problems of the elderly as they enter into an 1)aging society and utilize them to plan the overall plan of the city, compare the 2)lifestyle and patterns of the users and analyze the physical activity ability of the elderly. Specifically, the proposal to connect a single 4)garden path in the form of a 3)platform-type cohabitation is intended to try to resolve the isolation, depression, and social isolation of the elderly through a community between various users in the 3)platform and the 4)garden.
개선된 Padding기법을 이용한 Intra Prediction 성능 향상
김태영(Taeyoung Kim),이용빈(Yongbin Lee),이선율(Seonyul Lee),이선주(Seongjoo Lee) 대한전자공학회 2023 대한전자공학회 학술대회 Vol.2023 No.6
This paper analyzes prediction performance of intra prediction by combination of existing Padding technique (zero, half, mirror, circular, replication) and propose better Padding technique that can be used. Analyzed padding technique half & mirror, circular, replication, zero & mirror, circular, replication which is combination of two padding technique, and ‘Quartile padding’ using average. A simulation was performed on a grayscale 8-bit image using ‘visual studio 2019’. Prediction accuracy was evaluated by MSE (Mean Square Error) of the residual block.
움직임 감지를 위한 가우시안 혼합 모델 기반 전경탐지에서 발생하는 노이즈 처리 방법
정인범(In-Bum Chung),이용빈(Yongbin Lee),최동훈(Dong-Hoon Choi) 대한기계학회 2021 대한기계학회 춘추학술대회 Vol.2021 No.11
Detecting movements in a video feed can be done using either conventional computer vision techniques or deep learning. In this study, a foreground detection technique is used for movement detection in order to perform real-time analysis on a fixed camera. Foreground detection requires less computational resource compared to state-of-the-art machine-learning techniques, which makes it more suitable for real-time detection. Additionally, it has the benefit of not requiring the process of data labeling and model training, allowing detection of movements that were not anticipated by the engineer. However, there is a drawback concerning noisy predictions. In order to detect even the smallest changes, a sensitive detector has to be used. Unfortunately, this usually results more noise in detections. In this study, we attempt to find methods to sustain the sensitivity while removing circumstantial noise in the detections to enhance detection accuracy. This study focuses on the main causes of noises and methods to work around them.
다구치 방법과 근사최적설계를 이용한 자동차 연료탱크의 연료 넘침 방지 시스템 설계
박규병(Gyu-Byung Park),이용빈(Yongbin Lee),조인근(In-Geun Cho),최동훈(Dong-Hoon Choi) 대한기계학회 2013 大韓機械學會論文集A Vol.37 No.8
자동차 연료탱크는 크게 본체와 본체에 조립되는 부품들로 구성되어 있다. 본체에는 차량 주행시에 연료탱크에서 발생되는 증발가스를 배출하고, 연료가 외부로 유출되는 것을 방지하기 위한 여러 밸브들이 조립되어 있다. 하지만 이러한 밸브들은 연료 넘침에 주된 원인으로 알려져 있음에도 불구하고 현재 증발가스의 배출과 기구적인 위치만 고려하여 설계되고 있다. 따라서 본 연구에서는 밸브들의 기존기능을 유지하면서 연료 넘침을 최소화 시키기 위해 근사최적설계를 적용하였고, 다구치방법을 통해 실제 실험에서 근사최적설계의 유용성을 보였다. 결과적으로 최적화된 밸브 위치를 통해 개발기간과 비용을 절감하였고, 연료 넘침 최소화를 통해 자동차의 신뢰성을 향상시켰다. Automotive fuel tank is generally divided into two parts: main frame and assembly parts. While the car is running, valves are used to prevent liquid carry over and to discharge evaporated gas from the fuel tank. However, current fuel tank designs focus on the gas ventilation or secured location. In this study, the location of the parts used to prevent liquid carry over within the fuel tank is evaluated during an optimal design process. To develop this design process, an approximate optimization is applied. Through the optimal design process, the optimal valve location in fuel tank is determined and the approximate optimization is validated by the Taguchi method. Finally, the optimized valve location is used to reduce the development cost and time and to contribute toward improved automobile quality owing to enhanced reliability.
다층 퍼셉트론의 네트워크 아키텍쳐 결정을 위한 빅데이터 기반 효율적인 탐색 기법
류동흠(Dong Heum Ryu),이용빈(Yongbin Lee),최동훈(Dong-Hoon Choi) 대한기계학회 2020 대한기계학회 춘추학술대회 Vol.2020 No.12
A multilayer perceptron (MLP) is a deep learning model commonly used in various fields of engineering for building a regression model as a substitute for a nonlinear system. However, selection of MLP network architecture, which much influences the performance of an MLP, remains a challenge. In this study, an efficient search algorithm based on big data is proposed for selecting the appropriate network architecture. The proposed algorithm extracts 9 suitable candidates of network architectures and selects the structure with the best predictive performance for a new data. The candidates are extracted from a big data of approximately 1.4 million MLP network architectures optimized for benchmark regression problems with various tendencies. Various benchmark regression problems were used in order to verify the performance of the proposed algorithm and the results were compared to other existing algorithms.