http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
CARDB를 이용한 반복적인 업-다운 샘플링 네트워크 기반의 단일 영상 초해상도 복원
김인구(Ingu Kim),유송현(Songhyun Yu),정제창(Jechang Jeong) 한국방송·미디어공학회 2020 방송공학회논문지 Vol.25 No.2
Recently, many deep convolutional neural networks for image super-resolution have been studied. Existing deep learning-based super-resolution algorithms are architecture that up-samples the resolution at the end of the network. The post-upsampling architecture has an inefficient structure at large scaling factor result of predicting a lot of information for mapping from low-resolution to high-resolution at once. In this paper, we propose a single image super-resolution using Channel Attention Residual Dense Block based on an iterative up-down sampling architecture. The proposed algorithm efficiently predicts the mapping relationship between low-resolution and high-resolution, and shows up to 0.14㏈ performance improvement and enhanced subjective image quality compared to the existing algorithm at large scaling factor result.
잔차 블록 기반의 깊은 합성곱 신경망을 통한 단일 영상 초해상도 복원
김인구(Kim, Ingu),유송현(Yu, Songhyun),정제창(Jeong, Jaechang) 한국방송·미디어공학회 2018 한국방송공학회 학술발표대회 논문집 Vol.2018 No.11
신경망은 깊어질수록 gradient vanishing/exploding과 같은 네트워크가 불안정해지는 문제가 발생 한다. 잔차 블록을 이용하여 이러한 문제를 해결 할 수 있다. 본 논문에서는 영상 인식 분야에서 휼륭한 성능을 보여준 잔차 블록 기반의 깊은 합성곱 신경망을 통한 단일 영상 초해상도 복원 기법을 제안 한다. 제안한 알고리듬은 EDSR에 사용된 잔차 블록을 다양한 크기의 합성곱 연산을 통해 영상의 특징들을 다르게 분석하도록 수정하고 VDSR과 비숫한 수준의 복잡도로 구성하여 향상된 성능을 얻었다. 실험 결과, VDSR에 비해 PSNR이 최대 0.1㏈까지 증가했다.
CNG 機關의 水素混合率變化에 따른 性能및 稀薄燃燒特性에 관한 實驗的硏究
박승원(Seungwon Park),김인구(Ingu Kim),이성욱(Seangwock Lee),조용석(Yongseok Cho) 한국자동차공학회 2015 한국자동차공학회 학술대회 및 전시회 Vol.2015 No.11
Recently, the world faces the environmental problem such as air pollution due to harmful gas discharged from car and abnormal climate due to the green-house gases increased by the discharge of CO₂. So, the research on several fuels has been actively progressed throughout the world. Compressure Natural Gas(CNG), one of alternative for this problem, is less harmful, compared to the existing fossil fuel, as gaseous fuel, and less carbon in fuel ingredients and carbon dioxide generation rate relatively favorable more than the existing fuel. However, CNG fuel has the weakness of slow flame propagation speed and difficult fast burn. On the other hand, hydrogen does not include carbon in fuel ingredients, and does not discharge harmful gas such as CO and HC. Moreover, it has strength of quick burning velocity and ignition is possible with small ignition energy source and it`s has wide Lean Flammability Limit. If using this hydrogen with CNG fuel, the characteristics of output and discharge gas is improved by the mixer’s burning velocity improved, and, at the same time, is possible to have stable lean combustion with the reduction of CO₂ expected. Therefore, this research tries to identify the characteristics of engine and emission gas when mixing CNG fuel and hydrogen in each portion and burning them in spark igniting engine, and grasp the combustion stability and emission gas characteristics according to the change of access air ratio, and use it as the basic data of hydrogen-CNG premixed engine.
대량 주행 데이터 분석을 통한 소형 경유차의 NOx 배출특성 연구
이범호(Beomho Lee),윤보섭(Boseop Yun),정재우(Jaewoo Jung),김덕진(Deokjin Kim),차우섭(Woosub Cha),이상운(Sangwoon Lee),김인구(Ingu Kim) 한국자동차공학회 2018 한국 자동차공학회논문집 Vol.26 No.5
The introduction of the RDE(Real Driving Emissions) regulation of diesel vehicles has raised interest in emission characteristics on real roads and the operating characteristics of after-treatment systems, but there have been many restrictions on obtaining long-term driving and exhaust emission data in real road conditions that use PEMS (Portable Emission Measurement System). In this study, a SPEMS(Sensor-based Portable Emission Measurement System) was developed so it could efficiently acquire long-term driving and emission data by using the NOx sensor for an OBD system. It was installed in four vehicles that adhere to Euro 4-6 emission regulations to generate massive data under normal operating conditions in an actual road. The mean and MAW analysis results of the NOx emissions that were measured under long-term normal operating conditions have been found as highly correlated with the RDE test results performed on regulated driving routes with PEMS for certification. Furthermore, in the SCR applied to Euro-6 vehicles, the reduction efficiency of SCR decreased as the ambient temperature was about 5 °C or less, and the reduction efficiency was drastically lesser at the catalyst inlet temperature of 250 °C or less.