• The Journal of the Acoustical Society of Korea
• /
• v.39 no.6
• /
• pp.615-621
• /
• 2020

In this paper, we propose a multi-site based earthquake event classification method using graph convolution networks. In the traditional earthquake event classification methods using deep learning, they used single-site observation to estimate seismic event class. However, to achieve robust and accurate earthquake event classification on the seismic observation network, the method using the information from the multi-site observations is needed, instead of using only single-site data. Firstly, our proposed model employs convolution neural networks to extract informative embedding features from the single-site observation. Secondly, graph convolution networks are used to integrate the features from several stations. To evaluate our model, we explore the model structure and the number of stations for ablation study. Finally, our multi-site based model outperforms up to 10 % accuracy and event recall rate compared to single-site based model.

• Applied Chemistry for Engineering
• /
• v.10 no.3
• /
• pp.467-476
• /
• 1999

This paper analyzed the behavior of fixed-bed catalytic reactor (FBCR) which synthesizing maleic anhydride(MA) from the selective oxidation of n-butane. The behavior of FBCR describing convection-diffusion-reaction mechanism is examined by using two-dimensional pseudohomogeneous plug-flow transient model, with the kinetics of Langmuir-Hinshelwood type. Prediction model is composed by optimum parameter estimation from temperature profile, yield and conversion of single FBCR on operating condition variations of Sharma's pilot-plant experiment. A double FBCR with same yield and conversion for single FBCR generated a $8.96^{\circ}C$ lower hot spot temperature than a single FBCR. We could predict parametric sensitivity according to the variation of possible operating condition (temperature, concentration, volumetric flow of feed reactant and coolant flow rate) of single and double FBCR. Double FBCR showed the behavior of more operating range than single FBCR. Double FBCR with nonuniform activities could assure safety operation condition for the possible variation of operating condition. Also, double FBCR had slightly higher than the single FBCR in conversion and yield.

• Journal of Korean Society for Geospatial Information Science
• /
• v.15 no.2 s.40
• /
• pp.3-14
• /
• 2007

This study proposes an automatic method to generate 3D building models using a draft map, which is an intermediate product generated during the map generation process based on aerial photos. The proposed method is to generate a terrain model, roof models, and wall models sequentially from the limited 3D information extracted from an existing draft map. Based on the planar fitting error of the roof corner points, the roof model is generated as a single planar facet or a multiple planar structure. The first type is derived using a robust estimation method while the second type is constructed through segmentation and merging based on a triangular irregular network. Each edge of this roof model is then projected to the terrain model to create a wall facet. The experimental results from its application to real data indicates that the building models of various shapes in wide areas are successfully generated. The proposed method is evaluated to be an cost and time effective method since it utilizes the existing data.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.42 no.2
• /
• pp.108-118
• /
• 2014

On the optimization design problem using surrogate model, it requires considerable number of sampling points to construct a surrogate model which retains the accuracy. As an alternative to reduce construction cost of the surrogate model, Variable-Fidelity Modeling(VFM) technique, where correct high fidelity model based on the low fidelity surrogate model is introduced. In this study, hierarchical kriging model for variable-fidelity surrogate modeling is used and an optimization framework with multi-objective genetic algorithm(MOGA) is presented. To prove the feasibility of this framework, airfoil design optimization process is performed for the transonic region. The parameters of PARSEC are used to design variables and the optimization process is performed in case of varying number of grid and varying fidelity. The results showed that pareto front of all variable-fidelity models are similar with its single-level of fidelity model and calculation time is considerably reduced. Based on computational results, it is shown that VFM is a more efficient way and has an accuracy as high as that single-level of fidelity model optimization.

• The Journal of Korean Academy of Prosthodontics
• /
• v.57 no.2
• /
• pp.102-109
• /
• 2019

Purpose: The purpose of this study was to evaluate the accuracy of three types of intraoral scanners and the accuracy of the single abutment and bridge abutment model. Materials and methods: In this study, a single abutment, and a bridge abutment with missing first molar was fabricated and set as the reference model. The reference model was scanned with an industrial three-dimensional scanner and set as reference scan data. The reference model was scanned five times using the three intraoral scanners (CS3600, CS3500, and EZIS PO). This was set as the evaluation scan data. In the three-dimensional analysis (Geomagic control X), the divided abutment region was selected and analyzed to verify the scan accuracy of the abutment. Statistical analysis was performed using SPSS software (${\alpha}=.05$). The accuracy of intraoral scanners was compared using the Kruskal-Wallis test and post-test was performed using the Pairwise test. The accuracy difference between the single abutment model and the bridge abutment model was analyzed by the Mann-Whitney U test. Results: The accuracy according to the intraoral scanner was significantly different (P < .05). The trueness of the single abutment model and the bridge abutment model showed a statistically significant difference and showed better trueness in the single abutment (P < .05). There was no significant difference in the precision (P = .616). Conclusion: As a result of comparing the accuracy of single and bridge abutments, the error of abutment scan increased with increasing scan area, and the accuracy of bridge abutment model was clinically acceptable in three types of intraoral scanners.

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 1989.11a
• /
• pp.2-3
• /
• 1989

• Proceedings of the Korean Nuclear Society Conference
• /
• 2004.05a
• /
• pp.173.2-173.2
• /
• 2004

• Journal of Korean Society for Geospatial Information Science
• /
• v.17 no.3
• /
• pp.71-79
• /
• 2009

We expanded the 3D building information extraction method using shadow and vertical line from single high resolution image with meta information into the method for single high resolution image without meta information. Our method guesses an azimuth angle and an elevation angle of the sensor and the sun using reference building, selected by user, on an image. For test, we used an IKONOS image and an image extracted from the Google Earth. We calculated the Root Mean Square (RMS) error of heights extracted by our method using the building height extracted from stereo IKONOS image as reference, and the RMS error from the IKONOS image and the Google Earth image was under than 3 m. We also calculated the RMS error of horizontality position by comparison between building position extracted from only the IKONOS image and it from 1:1,000 digital map, and the result was under than 3 m. This test results showed that the height pattern of building models by our method was similar with it by the method using meta information.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.6
• /
• pp.536-543
• /
• 2023

Speaker diarization, which labels for "who spoken when?" in speech with multiple speakers, has been studied on a deep neural network-based end-to-end method for labeling on speech overlap and optimization of speaker diarization models. Most deep neural network-based end-to-end speaker diarization systems perform multi-label classification problem that predicts the labels of all speakers spoken in each frame of speech. However, the performance of the multi-label-based model varies greatly depending on what the threshold is set to. In this paper, it is studied a speaker diarization system using single-label classification so that speaker diarization can be performed without thresholds. The proposed model estimate labels from the output of the model by converting speaker labels into a single label. To consider speaker label permutations in the training, the proposed model is used a combination of Permutation Invariant Training (PIT) loss and cross-entropy loss. In addition, how to add the residual connection structures to model is studied for effective learning of speaker diarization models with deep structures. The experiment used the Librispech database to generate and use simulated noise data for two speakers. When compared with the proposed method and baseline model using the Diarization Error Rate (DER) performance the proposed method can be labeling without threshold, and it has improved performance by about 20.7 %.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11a
• /
• pp.115-118
• /
• 2002

소켓 프로그래밍(socket programming) 인터페이스를 지원하는 C/C++, perl, python 과 같은 언어들은 폴링(polling) 기능을 갖는 select() 함수를 제공한다. 이 select()함수를 이용할 경우, 단일 쓰레드(또는 프로세스)로 다중의 클라이언트 요청을 처리할 수 있다. 최근 네트워크 프로그래밍 분야에서 주목받는 자바 언어의 경우, 최신 JDK 1.4 의 비동기 입출력 패키지에서 select()함수를 제공하고 있으나, JDK 1.3을 포함한 그 이하의 버전에서는 아직까지 이 함수를 제공하지 않고 있다. 일반적으로 다중 쓰레드를 이용하여 소켓서버 응용프로그램을 개발할 경우, 코드가 단순해지고 응답이 빠른 장점이 있는 반면에 네트워크 연결이 증가할수록 다수의 쓰레드를 관리하는 일이 CPU에 큰 부담이 된다. 반면에 소켓폴링(socket polling)을 사용할 경우, 이러한 연결 유지에 대한 부담이 줄어드는 대신, 다중 쓰레드를 이용하는 방법에 비하여 구현이 어렵다. 본 논문에서는 다양한 시뮬레이션 환경에서 세가지 소켓 프로그래밍 모델에 대하여 그 성능을 비교평가 하였다. 이 세가지 모델은 단순 다중 쓰레드 모델(typical multi-thread model), 단일 쓰레드 소켓폴링 모델(socket polling with single-thread model), 다중 쓰레드 소켓폴링 모델(socket polling with multi-threadmodel)이다. 본 논문에서는 다중 쓰레드 소켓폴링 모델을 제안하고 JDK 1.3.1을 이용하여 구현하였다. 이 모델의 경우 복잡한 구조에도 불구하고 단순 다중 쓰레드 모델와 유사하거나 더 나은 성능을 보여주었다. 또한 동일한 용량의 쓰레드 풀(thread pool)을 사용하더라도 단순 다중 쓰레드 모델보다 더 많은 수의 클라이언트를 수용할 수 있는 장점이 있다. 이러한 결과를 바탕으로 본 연구팀에서 수행중인 MoIM-Messge서버의 네트워크 모듈로 다중 쓰레드 소켓폴링 모델을 적용하였다.