http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
석희억,한상술,김태중 경남대학교 환경문제연구소 2000 환경연구 Vol.23 No.-
This thesis investigates the styles of Modern Korean railway stations which constructed between 1899 and 1945. Railway stations' principal style of architecture can be divided into five different forms, which are Mixed Japanese Style with Western design element. Western Classic Style, Japanese Style incoporated Western in elevation, Korean Traditional Style and International Style. Mixed Japanese Style with Western design element usually shows the composition of rectangle plane with wooden-Japanese roof. Western Classic Style can be found prime locations railway stations such as Pusan, Shinuiju, and Seoul. In this case, Japan tries to spread the advanced-western architectural styles in the purpose of control the colonized country more efficiently. Japanese Style incorporated Western in elevation and plan in architectural preservation can be defined as the coexistence of Western and Mixed Japanese Style with Western design element architecture. Through 1930s, historical preservative and some of touristic cities' stations were developed as Korean Traditional Style. It constructed with concrete columns and brick walls instead of wooden structure and finished with mortar and materials and the style of openings and roof are influenced with Japanese style. Finally, in 1920s, it seems that the use of simple cubic shape of geometrical composition increases rather than using existing architectural styles. However, the stations in this period emphasize the cost-minimized functions of building as well as shorten the construction period rather than emphasizing the new western shape of architecture and the essence of modern architecture.
다중 임계 기법과 영역확장법을 이용한 뇌졸중 병변 정합 방법 연구
탁윤오,김태윤,황해길,김동억,최흥국 인제대학교 2009 仁濟論叢 Vol.24 No.1
In clinical studies of brain stroke, it generally involves the process that spatially normalizes MR images into the standard brain template in order to find out distributions of lesion. So far MNI305, which is people’s standard brain template, has been commonly used. However in this study, to reduce some morphological differences caused by human races, ages, and sexes, we used the Korean Statistical Probabilistic Anatomical Map(Korean-SPAM). As a preprocessing methods, brightness and position of a brain had been adjusted at the first Then, stroke lesions were segmented using multiple threshold and region growing methods. Finally, stroke lesions were registered to the Korean-SPAM using Mesh Warp algorithm. We can registered various stroke MR brain images to Korean-SPAM rightly by the proposed method.
Tae-Sun Yu,Hyun-Jung Kim,Tae-Eog Lee Institute of Electrical and Electronics Engineers 2017 IEEE transactions on semiconductor manufacturing Vol.30 No.2
<P>Wafer quality issues are becoming essential concerns in semiconductor manufacturing industry. It is becoming increasingly important for fab managers to raise the wafer quality level. Quality variation across wafers and wafer lots is also recognized as of vital importance. Wafer waiting times, which occur between consecutive wafer processing steps, are critical for the quality and quality variation of wafers. To resolve these quality issues, we consider waiting time constraints and variation in a flowshop. To follow the actual operating features of the fab, we define a two-machine flowshop with jobs that can skip the first process step and are ready to enter the second step from the beginning of scheduling. This research thus examines a machine scheduling problem that minimizes the variation in job waiting times in a generalized two-machine flowshop with skipping jobs and waiting time constraints. The mathematical properties of the problem such as the dominance properties and feasibility conditions are vigorously analyzed. These analyses provide profound insights into reduction of the search space in the solution procedure. We also observe that the derived properties are intuitively consistent with the well-known principles of queueing theory. From these, we develop efficient approximation algorithms and present their computational performance.</P>