자율주행환경 운동객체 탐지 및 추적 시스템

이병주(ByeongJu Lee),정지엽(Jiyeoup Jeong),최현준(Hyun Jun Choi),최진영(Jin Young Choi) 대한전자공학회 2018 대한전자공학회 학술대회 Vol.2018 No.6

This paper proposes an assistance system for autonomous vehicle which detects and tracks moving objects around the vehicle. A semantic segmentation network classifies objects and background, and makes a panorama-based motion detection algorithm suitable for autonomous driving environment. By fusing results of object detection and visual tracking, a final localization result of each object can be determined with spatio-temporal information. Through experiments, we show that the proposed method improve baseline algorithm, both of detection and tracking.

발효산삼배양근농축액의 산화방지 효과 및 과산화수소로 유발된 PC12 세포독성 보호효과

최여옥(Yeo Ok Choi),김유리(Yu-Ri Kim),신승용(Seung-Yong Shin),이재근(Jae Geun Lee),김철중(Chul Joong Kim),이예지(Ye ji Lee),강병주(Byeongju Kang),김관수(Gwansu Kim),최지은(Jee Eun Choi),한범석(Beom-Seok Han) 한국식품과학회 2018 한국식품과학회지 Vol.50 No.4

대부분의 신경퇴행성질환은 산화스트레스의 영향을 받는다. 따라서 본 연구에서는 진세노사이드 Rg5와 Rk1을 함유하고 있는 발효산삼배양근 농축액인 HLJG0701의 산화방지 효능을 확인하였다. 그 결과, HLJG0701는 효과적인 DPPH와 ABTS 라디칼제거능을 나타냈으며(IC<SUB>50</SUB>: 16배, 4배 희석액), 황산철(II) 유도군의 지방질 과산화물 발생 비교 시, 용량의존적으로 억제됨을 확인하였다(8배 희석액: 2.3 nM, 4배 희석액: 1.5 nM). HLJG0701의 세포 보호 효과를 MTT, LDH 분석으로 평가한 결과, 8, 16, 32, 64배 희석액의 HLJG0701를 처리한 경우 70, 53, 35, 26%의 세포생존율이 증가됨을 확인하였다. LDH 방출 측정 결과, H₂O₂로 인해 3.6배까지 증가한 LDH 활성은 8배 희석액의 HLJG0701 처리시 1.3배 수준까지 감소하였다. 또한 HLJG0701가 PC12세포에 발생한 산화스트레스 DCF-DA 분석(16배 희석액: 50% ROS 억제, 4배 희석액: 68% ROS 억제), TBARS (16배 희석액: 50.7% 감소, 4배 희석액: 46.5% 감소), GPx (16배 희석액: 133.3% 증가, 4배 희석액: 227.3% 증가), 그리고 SOD (16배 희석액: 118.2% 증가, 4배 희석액: 218.2% 증가) 분석으로 확인하였다. 위 결과는 HLJG0701가 산화방지 효능을 매개로 신경세포를 보호함으로써 신경퇴행성질환을 예방할 수 있는 가능성 있는 후보물질이 될 수 있다는 것을 의미하며, 임상시험 등 다양한 연구가 충분하지 않은 상태이므로 앞으로 지속적인 연구가 필요할 것으로 사료된다. Most neurodegenerative diseases are known to be influenced by oxidative stress. We investigated the antioxidative activity of the concentrate of fermented wild ginseng root culture (HLJG0701) containing ginsenosides Rg5 and Rk1. HLJG0701 showed effective DPPH and ABTS radical scavenging ability (IC<SUB>50</SUB>: 16- and 4-fold dilution, respectively) and was inhibited dose-dependently by the FeSO₄-induced lipid peroxidation group (8- and 4-fold dilution: 2.3 and 1.5 nM, respectively). In MTT and LDH assays, 8-, 16-, 32- and 64-fold diluted HLJG0701 significantly increased cell viability by 70, 53, 35, and 26%, respectively. LDH released by HLJG0701 was reduced 1.3-fold with 8-fold diluted HLJG0701 compared to the H₂O₂-treated control. In addition, the inhibitory effect of HLJG0701 on oxidative stress in PC12 cells was confirmed by DCF-DA analysis (16-, 4-fold diluted HLJG0701: 50 and 68% ROS inhibition, respectively), TBARS (16- and 4-fold diluted HLJG0701: 50.7 and 46.5% inhibition, respectively), GPx (16- and 4-fold diluted HLJG0701: 133.3 and 227.3% release, respectively), and SOD analysis (16- and 4-fold diluted HLJG0701: 118.2 and 218.2% release, respectively). These results suggested that HLJG0701 protects neuronal cells by its anti-oxidative effects and hence can be a potential preventive material against neurodegenerative diseases.

물체추적 시 오학습 방지를 위한 추적물체 가려짐 탐지

이병주(ByeongJu Lee),최진영(Jin Young Choi) 대한전자공학회 2016 대한전자공학회 학술대회 Vol.2016 No.6

신경망을 이용한 차량의 자율주행 전략 생성

이병주(ByeongJu Lee),허병호(Byeongho Heo),최진영(Jin Young Choi) 대한전자공학회 2017 대한전자공학회 학술대회 Vol.2017 No.6

This paper proposes a deep neural network generating an autonomous driving strategy using both reinforcement learning and supervised learning. To make the deep neural network learn lane keeping and lane changing. e develop a simulation environment of 3-lane road by modify a TORCS simulator. Following driver`s intention, the driving strategy network controls an autonomous vehicle`s steering. Through experiments, we show that the proposed method operates the autonomous vehicle to travel according to the driver`s intention, and the performance of the proposed method is trained successfully by reinforcement learning and supervised learning.

인휠모터 전기 자동차의 반능동 현가시스템 통합제어

김현동(Hyeandong Kim),최규재(Gyoojae Choi),하만호(Manho Ha),문홍주(Houngju Moon),강병주(Byeongju Kang) 한국자동차공학회 2015 한국자동차공학회 학술대회 및 전시회 Vol.2015 No.11

자율주행용 센서 공도 실증 방법 연구

이승민(Seungmin Lee),황윤형(Yunhyoung Hwang),강병주(Byeongju Kang),곽수진(Sujin Kwag),김성진(Sungjin Kim),최경수(Kyungsoo Choi) 한국자동차공학회 2020 한국자동차공학회 학술대회 및 전시회 Vol.2020 No.11

Carbon monoxide poisoning-induced encephalopathy in a carbon dioxide arc welder: a case report

Seongwon Ma,Hoekyeong Seo,Dong Joon Park,Byeongju Choi,Shinhee Ye 대한직업환경의학회 2022 대한직업환경의학회지 Vol.34 No.-

Background: It is widely known that carbon dioxide (CO₂) arc welding generates carbon monoxide (CO). However, to the best of our knowledge, no case reports have been published regarding CO poisoning in CO₂ arc welders. Therefore, we aimed to report a case of CO poisoning-induced encephalopathy in a CO₂ arc welder in the Republic of Korea to inform about the dangers of CO exposure among CO₂arc welders. Case presentation: A 40-year-old man working as a CO₂ arc welder for 15 years visited a local hospital with a tremor, involuntary urination, and speaking gibberish, on April 9, 2019. He stated that he had intermittent headache and forgetting symptoms for the last 5 years, and had been lost on the way to work several times. On April 9, 2019, he was diagnosed with CO poisoning-induced encephalopathy through brain magnetic resonance imaging. He received hyperbaric oxygen therapy, and some of his symptoms improved. According to the exposure assessment of his work environment, he was continuously exposed to high concentrations of CO for 15 years while operating CO₂ arc welding machines. Conclusions: After evaluating the patient’s work environment and evaluating his medical history, we concluded that his encephalopathy was caused by CO exposure during CO₂ arc welding. Thus CO₂ arc welders must be aware of the risk of CO poisoning and strive to avoid CO exposure.

악성코드 대응을 위한 신뢰할 수 있는 AI 프레임워크

신경아,이윤호,배병주,이수항,홍희주,최영진,이상진,Shin, Kyounga,Lee, Yunho,Bae, ByeongJu,Lee, Soohang,Hong, Heeju,Choi, Youngjin,Lee, Sangjin 한국정보보호학회 2022 정보보호학회논문지 Vol.32 No.5

Malware attacks become more prevalent in the hyper-connected society of the 4th industrial revolution. To respond to such malware, automation of malware detection using artificial intelligence technology is attracting attention as a new alternative. However, using artificial intelligence without collateral for its reliability poses greater risks and side effects. The EU and the United States are seeking ways to secure the reliability of artificial intelligence, and the government announced a reliable strategy for realizing artificial intelligence in 2021. The government's AI reliability has five attributes: Safety, Explainability, Transparency, Robustness and Fairness. We develop four elements of safety, explainable, transparent, and fairness, excluding robustness in the malware detection model. In particular, we demonstrated stable generalization performance, which is model accuracy, through the verification of external agencies, and developed focusing on explainability including transparency. The artificial intelligence model, of which learning is determined by changing data, requires life cycle management. As a result, demand for the MLops framework is increasing, which integrates data, model development, and service operations. EXE-executable malware and documented malware response services become data collector as well as service operation at the same time, and connect with data pipelines which obtain information for labeling and purification through external APIs. We have facilitated other security service associations or infrastructure scaling using cloud SaaS and standard APIs. 4차 산업혁명의 초연결사회에서 악성코드 공격은 더욱 기승을 부리고 있다. 이러한 악성코드 대응을 위해 인공지능기술을 이용한 악성코드 탐지 자동화는 새로운 대안으로 주목받고 있다. 그러나, 인공지능의 신뢰성에 대한 담보없이 인공지능을 활용하는 것은 더 큰 위험과 부작용을 초래한다. EU와 미국 등은 인공지능의 신뢰성 확보방안을 강구하고 있으며, 2021년 정부에서는 신뢰할 수 있는 인공지능 실현 전략을 발표했다. 정부의 인공지능 신뢰성에는 안전과 설명가능, 투명, 견고, 공정의 5가지 속성이 있다. 우리는 악성코드 탐지 모델에 견고를 제외한 안전과, 설명가능, 투명, 공정의 4가지 요소를 구현하였다. 특히 외부 기관의 검증을 통해 모델 정확도인 일반화 성능의 안정성을 입증하였고 투명을 포함한 설명가능에 중점을 두어 개발하였다. 변화무쌍한 데이터에 의해 학습이 결정되는 인공지능 모델은 생명주기 관리가 필요하다. 이에 인공지능 모델을 구성하는 데이터와 개발, 서비스 운영을 통합하는 MLOps 프레임워크에 대한 수요가 늘고 있다. EXE 실행형 악성코드와 문서형 악성코드 대응 서비스는 서비스 운영과 동시에 데이터 수집원이 되고, 외부 API를 통해 라벨링과 정제를 위한 정보를 가져오는 데이터 파이프라인과 연계하도록 구성하였다. 클라우드 SaaS 방식과 표준 API를 사용하여 다른 보안 서비스 연계나 인프라 확장을 용이하게 하였다.

초등학생의 성 성숙, 성 성숙 태도 및 신체 자존감

박선아,최은영,이은희,변미경,이병주,정계아,Park, Sunah,Choi, Eunyoung,Lee, Eunhee,Byeon, Mikyeong,Lee, Byeongju,Jung, Kyea 한국여성건강간호학회 2018 여성건강간호학회지 Vol.24 No.2

Purpose: The purpose of this study is to evaluate sexual maturation, attitudes toward sexual maturity, and body esteem in the sexual development of Korean elementary-school boys and girls. Methods: A descriptive cross-sectional study was conducted with 399 fifth and sixth graders (192 boys and 207 girls). The data were analysed with a $x^2$ test, t-test, and Pearson correlation coefficients. Results: Among the 207 girls, 70.5% had pubic hair growth, 68.1% had breast development, and 56.0% had a menstrual period. In boys, 59.4% of the 192 subjects experienced the development of external genitalia and 52.6% had pubic hair growth. Sexual maturation was related to grade (boys, t=7.07, p=.008; girls, t=12.76, p<.001), age (t=-2.20, p=.030; t=-4.11, p<.001), height (t=-5.16, p<.001; t=-7.52, p<.001), and weight (t=-2.89, p=.004; t=-5.19, p<.001) in both boys and girls. Girls were more likely to have sexual maturity than boys ($x^2=22.29$, p<.001). Boys showed more positive attitudes toward sexual maturity (t=2.10, p=.036) and higher body esteem (t=2.12, p=.035) than girls. Conclusion: This study shows that sexual maturation, attitude toward sexual maturity, and body esteem in sexual development differ between boys and girls. The findings indicate that it is necessary to develop a tailored sex-education program according to the sex of elementary-school children.