http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
박규태(GyuTae Park),박금비(GeumBi Park),Vasantha Kumar.C,고진환(JinHwan Koh) 한국통신학회 2021 한국통신학회 학술대회논문집 Vol.2021 No.11
인간과 기계간의 상호작용에서 인간의 손짓은 중요한 의미를 지니고 있다. 이러한 인간의 손짓은 점차 중요성이 증가하고 있지만, 복잡한 손짓의 입력이나 주변의 환경적 요인에 의한 잡음 등은 손짓 인식 기술 정확도 향상에 있어 해결해야 할 중요한 과제이다. 본 논문에서는 이러한 상황을 해결할 수 있는 기술로 Convolutional Neural Networks(CNN)을 제안한다. CNN은 이미지 데이터 학습에 유용하다는 장점을 가지고 있으며, 이 기술은 인간과 기계간의 상호작용에서 정확도를 매우 향상시킬 것이다. 5가지 수화동작을 7.4-9.0GHz의 주파수대역폭, 8dB의 이득 특성을 가진 Vivaldi안테나를 이용하여 데이터를 추출하였고, 전처리과정을 거친 데이터를 CNN을 통해 학습시켰다. 제안된 CNN의 분류 결과는 약 96%의 정확도를 보였다.
대규모 언어모델을 활용한 저자원 언어에 대한 Few-Shot 교차 언어 요약 기법
박규태(Gyutae Park),이환희(Hwanhee Lee) 대한전자공학회 2024 대한전자공학회 학술대회 Vol.2024 No.6
Cross-lingual summarization (XLS) aims to generate a summary in a target language different from the source language document. While large language models (LLMs) have shown promising zero-shot XLS performance, their few-shot capabilities on this task remain unexplored, especially for low-resource languages with limited parallel data. In this report, we investigate the few-shot XLS performance of various models, including fine-tuned mT5, GPT-3.5, GPT-4. Our findings highlight the potential of few-shot learning for improving XLS performance and the need for further research in designing LLM architectures and pre-training objectives tailored for this task. We recommend future work to explore more effective few-shot learning strategies and investigate the transfer learning capabilities of LLMs for cross-lingual summarization.
LSTM(Long Short Term Memory)을 이용한 RCS (Radar Cross Section)추정
박금비(Geumbi Park),바산타쿠마르(Vasantha Kumar C),박규태(Gyutae Park),고진환(Jinhwan Koh) 한국통신학회 2021 한국통신학회 학술대회논문집 Vol.2021 No.11
RCS측정은 전자기파의 산란과 반사가 중요한 역할을 하는 통신 시스템, 안테나 시스템 및 항공기 설계에 도움이 되는 요소이다. 그러나 항공기나 선박 등 대형 물체에서 RCS 측정 시간과 경비가 비교적 많이 든다. 따라서 본 논문에서는 앞의 문제를 해결하고, RCS측정 효율을 높이고자 LSTM의 방식을 도입하였다. 컴퓨터 시뮬레이션을 활용하여, 기존의 측정된 데이터로 허용 오차 범위내의 측정되지 않은 RCS 예측 데이터를 보여 주었다.
Measurement of Oil Thickness Using Snell’s Laser Refraction
Vasantha Kumar C(바산타쿠마르),Gyutae Park(박규태),Geumbi Park(박금비),Jinhwan Koh(고진환) 한국통신학회 2021 한국통신학회 학술대회논문집 Vol.2021 No.11
Concerns about oil spills are a major cause of marine contamination. If we can quantify the thickness of an oil slick, we will be able to better understand how oil spreads and behaves. Snell"s idea may be used to quantify refracted light beams in oil. A quantitative analysis of oil thickness was created using this method. A laser beam is focused towards the air/oil/water layer to conduct the test. When the refractive index of the medium changes, the beam rays refract. The location of a refracted laser beam ray is computed using Snell"s law, and the position is recorded by a camera module controlled remotely by a microcontroller. The X and Y coordinates of the beam ray location are determined by using Python OpenCV to analyze an image of the recorded beam position. The thickness of the oil is measured and evaluated after collecting the X and Y coordinates. The findings of the analysis were compared to the estimated thicknesses of the oil. The results of the tests show that the proposed method is accurate to within 1% error and can effectively quantify oil slick thicknesses on the water up to 11mm.
미립화와 환경 2 : 대형 경유트럭의 휘발성유화합물(VOCs) 배출특성 연구
문선희 ( Sunhee Mun ),이종태 ( Jihwan Son ),손지환 ( Gyutae Park ),박규태 ( Heungmin Yoo ),유흥민 ( Changwan Yun ),윤창완 ( Jeongsoo Kim ),김정수 ( Jongtae Lee ) 한국액체미립화학회 2015 한국액체미립화학회 학술강연회 논문집 Vol.2015 No.-
WHO(World Health Organization) IARC(International Agency for Research on Cancer) has issued to consider diesel engine exhaust gas as an one of HAPs which has carcinogenic for human(Group 1). In this study, it was conducted to evaluate variety characteristics on changing of VOCs by aftertreatment systems, and the heavy duty diesel trucks were chosen as experimental vehicles. The results of the VOCs emission characteristics according to aftertreatment systems showed that compared to vehicles equipped with DPF, the vehicles equipped with SCR had 12.3%, 15.1%, 27.8%, and 22.2% lower BTEX emissions of Benzene, Toluene, Ethylbenzene, and Xylene, respectively. The emissions per pollutants were in the order of Toluene > Xylene > Ethylbenzene > Benzene > Styrene, and Toluene had the highest levels of emissions at 16.25 mg/km and 13.80 mg/km, respectively, for the vehicles equipped with DPF and SCR. The results of emission characteristics from vehicle speed, when the vehicle was operated at low speeds below 10.6 km/h, showed that due to incomplete combustion, emission was 75% higher. The results of analyzing the correlation between THC and BTEX showed a proportionate correlation with the rate of change of the emissions. As a result of analyzing the BTEX/THC ratio(%) for each of vehicle speed, Toluene was shown to have the highest proportion, and BTEX took up about 19.1% of the THC.