• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.11a
• /
• pp.91-94
• /
• 2022

자율주행 자동차 개발 연구가 활발히 진행됨에 따라 객체 검출기의 성능이 중요하게 되었다. 딥러닝 기술의 발전하면서 객체 검출기의 성능도 큰 발전을 이루었다. 그에 따라 도로 위 차량 검출기의 성능도 발전하고 있으나 평상시 낮 도로상황에서 잘 동작하던 모델은 안개가 끼거나 밤 상황이 되면 제대로 동작하지 못하는 문제를 가지고 있다. 이유는 딥러닝 모델이 학습할 때 사용한 데이터셋의 정보에 따라 특정 도메인에 편향된 특성을 학습하기 때문이다. 따라서, 본 논문에서는 객체 검출 신경망에 도메인 판별기를 적용하여 이와 같은 도메인 이동 문제를 극복하는 모델을 제안한다. 모델의 성능을 Cityscapes 데이터셋과 Foggy Cityscapes 데이터셋을 사용하여 평가한 결과, 기존의 특정 도메인에서 학습한 모델보다 제안하는 모델의 검출 성능이 개선된다는 것을 확인하였다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.3
• /
• pp.341-346
• /
• 2022

Due to global warming and abnormal climate, the frequency and damage of floods are increasing, and the number of people exposed to flood-prone areas has increased by 25% compared to 2000. Floods cause huge financial and human losses, and in order to reduce the losses caused by floods, it is necessary to predict the flood in advance and decide to evacuate quickly. This paper proposes a flood risk determination model using a CNN-based classification model so that timely evacuation decisions can be made using rainfall and water level data, which are key data for flood prediction. By comparing the results of the CNN-based classification model proposed in this paper and the DNN-based classification model, it was confirmed that it showed better performance. Through this, it is considered that it can be used as an initial study to determine the risk of flooding, determine whether to evacuate, and make an evacuation decision at the optimal time.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.2
• /
• pp.465-470
• /
• 2024

The advancement of AI-based image generation technology has resulted in the creation of various images, emphasizing the need for technology capable of accurately discerning them. The amount of generated image data is limited, and to achieve high performance with a limited dataset, this study proposes a model for discriminating generated images using transfer learning. Applying pre-trained models from the ImageNet dataset directly to the CIFAKE input dataset, we reduce training time cost followed by adding three hidden layers and one output layer to fine-tune the model. The modeling results revealed an improvement in the performance of the model when adjusting the final layer. Using transfer learning and then adjusting layers close to the output layer, small image data-related accuracy issues can be reduced and generated images can be classified.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1996.11a
• /
• pp.833-840
• /
• 1996

중립면은 셸 (솔리드 모델) 유한 요소 생성, 로보트 이동 경로 계산, 특징 형상 판별 등에서 사용될 수 있다. 그러나 기존 중립면 계산 알고리즘들은 연립 방정식을 수렴성이 보장되지 않는 수치 해법으로 풀어야 했기 때문에 발전이 미비했다. 본 논문은 복셀-이등분 면의 교자성을 이용한 중립면 계산 알고리즘을 제시한다. 교차성은 보로노이 영역을 사용, 단순한 기하학적 요소간의 거리 비교로 판별한다. 이런 기하학적인 접근 방법은 기본적으로 수렴성 문제가 배제된다.

• Journal of Gifted/Talented Education
• /
• v.18 no.3
• /
• pp.371-400
• /
• 2008

The purposes of this study were to develop and validate Korean MHBT for identification of giftedness. MHBT in this study consists of KFT-HB and MHBT-5. MHBT-S was composed of 1) space presentation and thinking ability, space perception,. physics/technic tasks 2) affective domain; creativity, achievement motivation, desire of knowledge, social competence questionnaire 3) performance attitude questionnaire 4) interest questionnaire. The subject were 489 middle school students (1 or 2grade) in the education centers for gifted youth and general classes. Except a few subscales, internal consistent reliability was considered good. Korean MHBT discriminated well gifted students from general students in KFT-HB and some subtests of MHBT-5. As results, Korean MHBT in this study was expected to be a reliable and valid instrument for identification of korean gifted students.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.12 no.4
• /
• pp.19-33
• /
• 2008

Dynamic response measurements from natural excitation were carried out for 25- and 42-story buildings to evaluate their inherent properties, such as natural frequencies, mode shapes and damping ratios. Both are reinforced concrete buildings adopting a core wall, or with shear walls as the major lateral force resisting system, but frames are added in the plan or elevation. In particular, shear walls in a 25-story building are converted to frames from the 4th floor level downwards while maintaining a core wall throughout, resulting in a fairly complex structure. Due to this, along with similar stiffness characteristics in the principal directions, significantly coupled and closely spaced modes of motion are expected in this building, making identification rather difficult. By using various state-of-the-art system identification methods, the modal parameters are extracted, and the results are then compared. Three frequency-domain and four time-domain based operational modal identification methods are considered. Overall, all natural frequencies and damping ratios estimated from the different identification methods showed a greater consistency for both buildings, while mode shapes exhibited some degree of discrepancy, varying from method to method. On the other hand, in comparison with analysis results obtained using the initial finite element(FE) models, test results exhibited a significant difference of about doubled frequencies, at least for the three lower modes in both buildings. To improve the correlation between test and analysis, a few manual schemes of FE model updating based on plausible reasons have been applied, and acceptable results are obtained. The advantages and disadvantages of each identification method used are addressed, and some difficulties that might arise from the updating of FE models, including automatic procedures, for such large structures are carefully discussed.

• Analytical Science and Technology
• /
• v.32 no.1
• /
• pp.7-16
• /
• 2019

The objective of this study was the development of a discrimination model for the cultivational origin of paper mulberry bast fiber and Hanji using near infrared (NIR) and mid infrared (MIR) spectroscopy combined with partial least squares discriminant analysis (PLS-DA). Paper mulberry bast fiber was purchased in 10 different regions of Korea, and used to make Hanji. PLS-DA was performed using pre-treated FT-NIR and FT-MIR spectral data for paper mulberry bast fiber and Hanji. PLS-DA of paper mulberry bast fiber and Hanji samples, using FT-NIR spectral data, showed 100 % performance in cross validation and the confusion matrix (accuracy, sensitivity, and specificity). The discrimination models showed four regional groups which demonstrated clearer separation and much superior score plots in the NIR spectral data-based model than in the MIR spectral data-based model. Furthermore, the discrimination model based on the NIR spectral data of paper mulberry bast fiber had highly similar score morphology to that of the discrimination model based on the NIR spectral data of Hanji.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.645-648
• /
• 2007

IT 에 대한 비즈니스 의존성이 증가하면서, 안정된 IT 서비스의 제공과 비용 효과적인 운영의 중요성이 강조되고 있다. ITIL 에서는 효율적이고 신속한 변경 처리를 위해 Service Support 영역에서 변경 관리 프로세스를 제시하고 있다. 하지만 고비용을 요하는 CAB 의 소집 여부 등 의사 결정이 변경 관리자의 자의적인 판단에 의존함으로써, 다른 비즈니스 및 안정된 IT 서비스 제공의 위험요소로 작용하고 있다. 본 논문에서는 이러한 문제점을 해결하기 위하여 요청된 변경 사안이 신중한 검토가 필요한 지 여부를 객관적으로 판단할 수 있도록, 판별분석기법을 적용한 변경 모델 결정 방법을 제안한다. 제안된 모델의 유효성을 검증하기 위해, 실제 운영에 적용된 변경 관리 모델과 제안된 모델을 이용했을 때의 결과를 비교하고 그 결과를 제시하였다. 제안된 방법은 동일한 사안에 대하여 일관성 있는 결정을 도출할 수 있어 프로세스 품질개선에 기여할 수 있으며, 궁극적으로 안정된 IT 서비스 제공에 기여하여 기업성과를 개선할 수 있을 것이다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.6
• /
• pp.1010-1015
• /
• 2008

The choice of RFID system is recently progressing(being) rapidly at various field. For the sake of RFID system popularization, However, We should solve privacy invasion to gain the pirated information of RFID tag. There is the safest M Ohkubos's skill among preexistent studying to solve these problems. But, this skill has a problem that demands a immense calculation capability caused an increase in tag number when we discriminate tags. So, This paper proposes the way of transplant to Grid environment for keeping Privacy Protection up and reducing the Tag Identification Time. And, We propose the Tag Identification Process Model to apply Even Division Algorithm to separate SP with same site in each node. If the proposed model works in Grid environment at once, it would reduce the time to identify tags to 1/k.

• Journal of IKEEE
• /
• v.26 no.3
• /
• pp.494-499
• /
• 2022

It is reported that genome-wide RNA-seq profiles has potential as biomarkers of aging. A number of researches achieved promising prediction performance based on gene expression profiles. We develop an age prediction method based on the transcriptome of human dermal fibroblasts by selecting a proper age interval. The proposed method executes multiple rules in a sequential manner and a rule utilizes a classifier and a regression model to determine whether a given test sample belongs to the target age interval of the rule. If a given test sample satisfies the selection condition of a rule, age is predicted from the associated target age interval. Our method predicts age to a mean absolute error of 5.7 years. Our method outperforms prior best performance of mean absolute error of 7.7 years achieved by an ensemble based prediction method. We observe that it is possible to predict age based on genome-wide RNA-seq profiles but prediction performance is not stable but varying with age.