• 반도체디스플레이기술학회지
• /
• 제23권2호
• /
• pp.38-44
• /
• 2024

This research proposed DBSCAN-D, which is a clustering technique for locating POI based on existing density-based clustering research, such as GPS data, generated by moving objects. This method is designed based on 'staying time' and 'directionality' extracted from the relationship between GPS data. The staying time can be extracted through the difference in the reception time between data using the time at which the GPS data is received. Directionality can be expressed by moving the area of data generated later in the direction of the position of the previously generated data by concentrating on the point where the GPS data is sequentially generated. Through these two properties, it is possible to perform clustering suitable for the data set generated by the moving object.

• 천문학회보
• /
• 제46권1호
• /
• pp.41.3-42
• /
• 2021

Conventional global magnetic field maps, such as daily updated synoptic maps, have been constructed by merging together a series of observations from the Earth's viewing direction taken over a 27-day solar rotation period to represent the full surface of the Sun. It has limitations to predict real-time farside magnetic fields, especially for rapid changes in magnetic fields by flux emergence or disappearance. Here, we construct accurate synchronic magnetic field maps using frontside and AI-generated farside data. To generate the farside data, we train and evaluate our deep learning model with frontside SDO observations. We use an improved version of Pix2PixHD with a new objective function and a new configuration of the model input data. We compute correlation coefficients between real magnetograms and AI-generated ones for test data sets. Then we demonstrate that our model better generate magnetic field distributions than before. We compare AI-generated farside data with those predicted by the magnetic flux transport model. Finally, we assimilate our AI-generated farside magnetograms into the flux transport model and show several successive global magnetic field data from our new methodology.

• 한국측량학회지
• /
• 제33권3호
• /
• pp.155-161
• /
• 2015

Dike lines are important features for describing the detailed shapes of dikes and for detecting topographic changes on dike surfaces. Historically, dike lines have been generated using only the LiDAR data. This paper proposes a new methodology for selecting an appropriate dike line on various dike surfaces using the topographic LiDAR data and multispectral orthoimages taken in the Nakdong River basins. The fi rst baselines were generated from the given LiDAR data using the modified convex hull algorithm and smoothing spline function, and the second baselines were generated from the given orthoimages by the Canny operator. Next, one baseline was selected among the two baselines at 10m intervals by comparing their elevations, and the selected baseline at 10m interval was defined as the dike line segment. Finally, the selected dike line segments were connected to construct the 3D dike lines. The statistical results show that the dike lines generated using both the LiDAR data and multispectral orthoimages had the improved horizontal and vertical accuracies than the dike lines generated only using the LiDAR data on the various dike surfaces.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2009년도 학술발표회 초록집
• /
• pp.1211-1214
• /
• 2009

The goal of this research is to apply the neural networks models for the disaggregation of the pan evaporation (PE) data, Republic of Korea. The neural networks models consist of generalized regression neural networks model (GRNNM) and multilayer perceptron neural networks model (MLP-NNM), respectively. The disaggregation means that the yearly PE data divides into the monthly PE data. And, for the performances of the neural networks models, they are composed of training and test performances, respectively. The training data consist of the generated data using PARMA (1,1). And, the testing data consist of the historic data, respectively. From this research, we evaluate the impact of GRNNM and MLP-NNM for the disaggregation of the nonlinear time series data. We should, furthermore, construct the credible data of the monthly PE data from the disaggregation of the yearly PE data, and can suggest the methodology for the irrigation and drainage networks system.

• Journal of information and communication convergence engineering
• /
• 제19권4호
• /
• pp.228-233
• /
• 2021

In this study, we analyze the credit information (loan, delinquency information, etc.) of individual business owners to generate voluminous training data to establish a bankruptcy prediction model through a partial synthetic training technique. Furthermore, we evaluate the prediction performance of the newly generated data compared to the actual data. When using conditional tabular generative adversarial networks (CTGAN)-based training data generated by the experimental results (a logistic regression task), the recall is improved by 1.75 times compared to that obtained using the actual data. The probability that both the actual and generated data are sampled over an identical distribution is verified to be much higher than 80%. Providing artificial intelligence training data through data synthesis in the fields of credit rating and default risk prediction of individual businesses, which have not been relatively active in research, promotes further in-depth research efforts focused on utilizing such methods.

• International Journal of Advanced Culture Technology
• /
• 제9권3호
• /
• pp.327-333
• /
• 2021

Inbody is a tool for measuring health information with high reliability and accuracy to analyze body composition. Unlike the existing method of storing/processing and outputting data on the server side, the health information generated by InBody requires accurate support for health sharing and data analysis services using mobile devices. However, in the process of transmitting body composition measurement information to a mobile service, a problem may occur in data transmission/reception processing. The reason for this is that, since the network network in the cloud environment is used, if the connection is cut off or the connection is changed, it is necessary to provide a global service, not a temporary area, focusing on the mobility of InBody information. In addition, since InBody information is transmitted to mobile devices, a standard schema should be defined in the mobile cloud environment to enable information transfer between standardized InBody data and mobile devices. We propose a mobile cloud system using EMRA(Extended Metadata Registry Access) in which a mobile device processes and transmits body data generated in the inbody and manages the data of each local organization with a standard schema. The proposed system processes the data generated in InBody and converts it into a standard schema using EMRA so that standardized data can be transmitted. In addition, even when the mobile device moves through the area, the coordinator subsystem is in charge of providing access services. In addition, EMRA is applied to the collision problem due to schema heterogeneity occurring in the process of accessing data generated in InBody.

• Journal of Information Display
• /
• 제10권1호
• /
• pp.6-15
• /
• 2009

In this paper, a new approach is proposed for the efficient generation of computer-generated holograms (CGHs) using the spatially redundant data on a 3D object and the novel look-up table (N-LUT) method. First, the pre-calculated N-point principle fringe patterns (PFPs) were calculated using the 1-point PFP of the N-LUT. Second, spatially redundant data on a 3D object were extracted and re-grouped into the N-point redundancy map using the run-length encoding (RLE) method. Then CGH patterns were generated using the spatial redundancy map and the N-LUT method. Finally, the generated hologram patterns were reconstructed. In this approach, the object points that were involved in the calculation of the CGH patterns were dramatically reduced, due to which the computational speed was increased. Some experiments with a test 3D object were carried out and the results were compared with those of conventional methods.

• 대한토목학회논문집
• /
• 제11권4호
• /
• pp.113-119
• /
• 1991

본 연구에서는 월 강우량과 월 증발량 자료만 있는 하천유역에 대하여 장기 물수지 분석을 실시하는 방법론을 제시하고져 하였다. 단기간의 월 강우량 자료를 경혐공식에 의해 월 유출량 자료로 변환시킨 후 추계학적 모의발생 모형을 사용하여 이들 단기 유출자료로부터 일군의 장기 유출자료계열을 발생시켰고, 자료계열별로 갈수빈도해석에 의해 최대 갈수기간 및 월 강수량계열을 작성하였다. 계획년도별 각종 용수수요를 표준절차에 의해 추정하였으며 순 물소모량도 계산하였다. 유역내의 기존 저수지를 총괄하는 합성저수지를 통해 Deficit-Supply 방법으로 물 수지분석을 실시한 결과 물 부족량은 갈수재현기간이 커짐에 따라 급격하게 커지는 것으로 나타났다. 이는 하천 유역의 장기 물 수지분석을 통해 신뢰성있는 물 부족량을 계산하기 위해서는 추계학적 모의발생모형에 의한 장기간 유출량의 발생이 필수적이며 수자원 시스템의 적정 갈수재현기간의 선정이 대단히 중요함을 시사해 주는 것이다.

• 산업기술연구
• /
• 제20권A호
• /
• pp.131-138
• /
• 2000

A past monthly data is not faithful so much for a short term. But, the stochastic generation technique was provide of a long-term data. Thus this study is used a data which generated a monthly inflow amounts data by Thomas-Fiering model. This model is needed a certain process which determination of distribution, decision of continuous durability, etc. It was generated a inflow data every one month as Thomas-Fiering method. The generated inflow data was used input data for a monthly cumulative analysis. This analysis obtained a storage capacities which would be required during droughts having various return periods. It was presented a equation of fitting regression that was carried out regression analysis of 5, 10, 20, 50 years period.

• Journal of Web Engineering
• /
• 제21권2호
• /
• pp.497-522
• /
• 2021

Audio event classification refers to the detection and classification of non-verbal signals, such as dog and horn sounds included in audio data, by a computer. Recently, deep neural network technology has been applied to audio event classification, exhibiting higher performance when compared to existing models. Among them, a convolutional neural network (CNN)-based training method that receives audio in the form of a spectrogram, which is a two-dimensional image, has been widely used. However, audio event classification has poor performance on test data when it is recorded by a device (unknown device) different from that used to record training data (known device). This is because the frequency range emphasized is different for each device used during recording, and the shapes of the resulting spectrograms generated by known devices and those generated by unknown devices differ. In this study, to improve the performance of the event classification system, a CNN based on the log mel-spectrogram separation technique was applied to the event classification system, and the performance of unknown devices was evaluated. The system can classify 16 types of audio signals. It receives audio data at 0.4-s length, and measures the accuracy of test data generated from unknown devices with a model trained via training data generated from known devices. The experiment showed that the performance compared to the baseline exhibited a relative improvement of up to 37.33%, from 63.63% to 73.33% based on Google Pixel, and from 47.42% to 65.12% based on the LG V50.