http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
다중 스케일 어텐션과 심층 앙상블 기반 동물 피부 병변 분류 기법
곽민호,김경태,최재영,Kwak, Min Ho,Kim, Kyeong Tae,Choi, Jae Young 한국멀티미디어학회 2022 멀티미디어학회논문지 Vol.25 No.8
Skin lesions are common diseases that range from skin rashes to skin cancer, which can lead to death. Note that early diagnosis of skin diseases can be important because early diagnosis of skin diseases considerably can reduce the course of treatment and the harmful effect of the disease. Recently, the development of computer-aided diagnosis (CAD) systems based on artificial intelligence has been actively made for the early diagnosis of skin diseases. In a typical CAD system, the accurate classification of skin lesion types is of great importance for improving the diagnosis performance. Motivated by this, we propose a novel deep ensemble classification with multi-scale attention networks. The proposed deep ensemble networks are jointly trained using a single loss function in an end-to-end manner. In addition, the proposed deep ensemble network is equipped with a multi-scale attention mechanism and segmentation information of the original skin input image, which improves the classification performance. To demonstrate our method, the publicly available human skin disease dataset (HAM 10000) and the private animal skin lesion dataset were used for the evaluation. Experiment results showed that the proposed methods can achieve 97.8% and 81% accuracy on each HAM10000 and animal skin lesion dataset. This research work would be useful for developing a more reliable CAD system which helps doctors early diagnose skin diseases.
Finding Best Model in Multiple Regression Applying Reversible Jump MCMC
곽민호,김미림 한국자료분석학회 2019 Journal of the Korean Data Analysis Society Vol.21 No.4
The purpose of this study was to demonstrate reversible jump MCMC (RJ-MCMC) in multiple regression. The RJ-MCMC determined the best number of predictors given the data. Specifically, in the empirical analysis, the results showed that 12 predictors were selected as the appropriate predictors to explain science achievement among 20 predictors. It seems that the RJ-MCMC prefer to a simpler model compared to the other model selection methods (i.e., AIC, forward selection, backward selection, and stepwise selection). However, BIC suggested the same number of the variables suggested by the RJ-MCMC. To compare the model selection based on BIC and the RJ-MCMC, the simulation study was performed. The general trend for both model selection results is that the accuracy and the precision of model selection improves when the sample size is large or the number of the predictors is small. However, the model selection via the RJ-MCMC shows the better performance than the performance based on BIC when the number of predictors increases. Also, the results might imply that the RJ-MCMC allows to select the variable even the magnitude of the variables is small with parsimoniousness.
대차 형상 변화에 의한 차세대 고속열차 항력 특성에 대한 실험적 연구
곽민호,이영빈,이정욱,김규홍,이동호,정형석,장영일,권혁빈 한국항공우주학회 2011 한국항공우주학회 학술발표회 논문집 Vol.2011 No.4
현재 개발 중인 차세대 고속열차의 대차 형상 변화에 의한 항력 특성을 알아보기 위해 풍동실험을 수행하였다. 1/20 스케일의 고속열차 3량 모델을 이용하여 속도 30,40,50,60㎧에서 대차 형상을 일반대차, 대차커버, 유선형의 세가지를 조합하여 실험을 수행하였다. 일반 대차에 비해 대차 커버에 의한 항력 저감 효과보다 유선형에 의한 항력 저감 효과가 더 크게 나타났다. 위치별로는 유선형 형상과 대차커버 형상 모두 전두부 1번 대차에서 저감 효과가 가장 크게 나타났고 유선형 형상의 경우 전두부 2번 대차에서, 대차 커버 형상의 경우 후미부 2번 대차에서 저감 효과가 가장 작게 나타났다. Wind tunnel tests are performed in order to analyze the aerodynamic drag characteristics of HEMU-400x, next generation high speed train under development. With 1/20 scaled 3-car model, the experiments are done at 30, 40, 50, 60㎧ using a normal bogie, a bogie cover, and a streamlined shape. Comparing to the aerodynamic drag with a normal bogie, the aerodynamic drag reduction by a streamlined shape is much than by a bogie cover. At 1st bogie cover of 1st car, the reduction is most both by a streamlined shape and by a bogie cover. The reduction is least at the 2nd bogie cover of 1st car by a streamlined shape and at the 2nd bogie cover of 3rd car by a bogie cover.
공력 하중 저감을 위한 고속철도차량 전두부 해치 커버의 최적 설계
곽민호,강형민,윤수환,권태수 한국전산유체공학회 2023 한국전산유체공학회지 Vol.28 No.3
Design optimization of frontal hatch cover of high speed train was preformed for reduction of aerodynamic load. For this, the shape function was used to model the frontal hatch cover. Then, a total of 15 design points were selected using three design variables: the length of the hatch cover, top-view curvature, and side-view curvature. A three-dimensional flow analysis was performed for each design point. Finally, the optimal shape was derived through optimization, and the aerodynamic load of the optimal shape was compared with that of baseline model. As a result, the length of the frontal hatch increased compared to the baseline model and the aerodynamic load was improved by about 20% from 4,890N to 3,920N through the optimized shape.
A comparison between logistic regression and neural networks in a constructed response item study
곽민호,박철우 한국데이터정보과학회 2019 한국데이터정보과학회지 Vol.30 No.5
The purpose of the study is to demonstrate the prediction quality of logistic regression and artificial neural networks. The main results of the study are the comparisons of the accuracy of both methods. The response variable of the model is a comment assignment by a human rater, and the four predictors are topic proportions estimated from latent Dirichlet allocation. The constructed models for both analyses are mainly concerned with predicting the comment assignment by using the topic proportions as the predictors. The results show that the accuracy of the test data set is generally higher than the accuracy of the cross-validation quality of the logistic regression, and these results are well matched with previous empirical studies. Also, although the use of this accuracy for practical purposes remains still questionable, the results reveal the potential utility the neural network if larger sample size is available in the future.