• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.8-8
• /
• 2015

기후변화에 따른 집중호우 및 태풍 발생의 증가로 강우레이더를 이용한 홍수예경보시스템의 필요성이 증대되고 있다. 그러나 현재 국내에서 주로 활용되고 있는 단일편파 레이더는 정확도의 한계로 인해 홍수예보 활용에 어려움을 야기해왔다. 최근에는 수직반사도, 차등반사도, 비차등반사도 등 다양한 변수 취득을 통해 강우입자의 형태를 더욱 정확하게 추정할 수 있는 이중편파 레이더의 활용이 높아지고 있다. 본 연구에서는 홍수예보 활용을 위해 이중편파 레이더 실황강우 및 예측강우의 정확도를 평가하고자 한다. 평가를 위해 비슬산 레이더 자료를 활용하였으며, 2012~2014년의 강우사상을 선정하였다. 단일 및 이중편파 레이더 강우를 각각 추정하고, 강우예측을 위해 추정된 레이더 강우를 이류모델(Translation model)에 연계하여 선행 6시간까지의 예측강우를 생산하였다. 강우의 탐지능력 평가를 위해 Hit rate를 이용하였으며, 레이더 관측반경 증가 및 강우강도의 증가에 따른 정확도 분석을 수행하였다. 강수추정 정확도 평가를 위해 상관계수와 평균제곱근 오차를 이용하였으며, 비슬산 강우레이더 100 km 반경 내에 속한 국토교통부 관할의 지상관측강우와비교하였다. 그 결과, 이중편파 레이더 실황강우가 단일편파 레이더에 비해 지상관측강우의 거동과 더욱 유사하게 나타났으며, 양적인 오차도 더 적은 것으로 확인되었다. 또한, 레이더 예측강우는 선행시간이 증가함에 따라 정확도가 감소하였으나, 선행시간 1시간까지는 활용이 가능하다고 판단된다.

• Fibers and Polymers
• /
• v.7 no.3
• /
• pp.310-316
• /
• 2006

In this article, an attempt has been made to explain the failure mechanism of spun yams. The mechanism includes the aspects of generation and distribution of forces on a fibre under the tensile loading of a yam, the free body diagram of forces, the conditions for gripping and slipping of a fibre, and the initiation, propagation, and ultimate yam rupture in its weakest link. A simple mathematical model for the tenacity of spun yams has been proposed. The model is based on the translation of fibre bundle tenacity into the yam tenacity.

• Journal of Korea Water Resources Association
• /
• v.53 no.12
• /
• pp.1159-1172
• /
• 2020

This study applied deep convolution neural network based on U-Net and SegNet using long period weather radar data to very short-term rainfall prediction. And the results were compared and evaluated with the translation model. For training and validation of deep neural network, Mt. Gwanak and Mt. Gwangdeoksan radar data were collected from 2010 to 2016 and converted to a gray-scale image file in an HDF5 format with a 1km spatial resolution. The deep neural network model was trained to predict precipitation after 10 minutes by using the four consecutive radar image data, and the recursive method of repeating forecasts was applied to carry out lead time 60 minutes with the pretrained deep neural network model. To evaluate the performance of deep neural network prediction model, 24 rain cases in 2017 were forecast for rainfall up to 60 minutes in advance. As a result of evaluating the predicted performance by calculating the mean absolute error (MAE) and critical success index (CSI) at the threshold of 0.1, 1, and 5 mm/hr, the deep neural network model showed better performance in the case of rainfall threshold of 0.1, 1 mm/hr in terms of MAE, and showed better performance than the translation model for lead time 50 minutes in terms of CSI. In particular, although the deep neural network prediction model performed generally better than the translation model for weak rainfall of 5 mm/hr or less, the deep neural network prediction model had limitations in predicting distinct precipitation characteristics of high intensity as a result of the evaluation of threshold of 5 mm/hr. The longer lead time, the spatial smoothness increase with lead time thereby reducing the accuracy of rainfall prediction The translation model turned out to be superior in predicting the exceedance of higher intensity thresholds (> 5 mm/hr) because it preserves distinct precipitation characteristics, but the rainfall position tends to shift incorrectly. This study are expected to be helpful for the improvement of radar rainfall prediction model using deep neural networks in the future. In addition, the massive weather radar data established in this study will be provided through open repositories for future use in subsequent studies.

• Journal of KIISE:Software and Applications
• /
• v.32 no.5
• /
• pp.385-395
• /
• 2005

Long sentence analysis has been a critical problem in machine translation because of high complexity. The methods of intra-sentence segmentation have been proposed to reduce parsing complexity. This paper presents the intra-sentence segmentation method based on maximum entropy probability model to increase the coverage and accuracy of the segmentation. We construct the rules for choosing candidate segmentation positions by a teaming method using the lexical context of the words tagged as segmentation position. We also generate the model that gives probability value to each candidate segmentation positions. The lexical contexts are extracted from the corpus tagged with segmentation positions and are incorporated into the probability model. We construct training data using the sentences from Wall Street Journal and experiment the intra-sentence segmentation on the sentences from four different domains. The experiments show about $88\%$ accuracy and about $98\%$ coverage of the segmentation. Also, the proposed method results in parsing efficiency improvement by 4.8 times in speed and 3.6 times in space.

• Electronics and Telecommunications Trends
• /
• v.33 no.4
• /
• pp.1-13
• /
• 2018

Deep learning (DL) models have been widely applied to AI applications such image recognition and language translation with big data. Recently, DL models have becomes larger and more complicated, and have merged together. For the accelerated training of a large-scale deep learning model, model parallelism that partitions the model parameters for non-shared parallel access and updates across multiple machines was provided by a few distributed deep learning frameworks. Model parallelism as a training acceleration method, however, is not as commonly used as data parallelism owing to the difficulty of efficient model parallelism. This paper provides a comprehensive survey of the state of the art in model parallelism by comparing the implementation technologies in several deep learning frameworks that support model parallelism, and suggests a future research directions for improving model parallelism technology.

• The Journal of Advanced Prosthodontics
• /
• v.9 no.4
• /
• pp.244-251
• /
• 2017

PURPOSE. The purpose of this study was to determine fracture resistance and failure modes of three-unit fixed dental prostheses (FDPs) made of lithium disilicate pressed on zirconia (LZ), monolithic lithium disilicate (ML), and monolithic zirconia (MZ). MATERIALS AND METHODS. Co-Cr alloy three-unit metal FDPs model with maxillary first premolar and first molar abutments was fabricated. Three different FDPs groups, LZ, ML, and MZ, were prepared (n = 5 per group). The three-unit FDPs designs were identical for all specimens and cemented with resin cement on the prepared metal model. The region of pontic in FDPs was given 50,000 times of cyclic preloading at 2 Hz via dental chewing simulator and received a static load until fracture with universal testing machine fixed at $10^{\circ}$. The fracture resistance and mode of failure were recorded. Statistical analyses were performed using the Kruskal-Wallis test and Mann-Whitney U test with Bonferroni's correction (${\alpha}=0.05/3=0.017$). RESULTS. A significant difference in fracture resistance was found between LZ ($4943.87{\pm}1243.70N$) and ML ($2872.61{\pm}658.78N$) groups, as well as between ML and MZ ($4948.02{\pm}974.51N$) groups (P<.05), but no significant difference was found between LZ and MZ groups (P>.05). With regard to fracture pattern, there were three cases of veneer chipping and two interfacial fractures in LZ group, and complete fracture was observed in all the specimens of ML and MZ groups. CONCLUSION. Compared to monolithic lithium disilicate FDPs, monolithic zirconia FDPs and lithium disilicate glass ceramics pressed on zirconia-based FDPs showed superior fracture resistance while they manifested comparable fracture resistances.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.2 s.308
• /
• pp.81-92
• /
• 2006

This paper proposes a robustness watermarking for 3D mesh model based on projection onto convex sets (POCS). After designing the convex sets for robustness and invisibility among some requirements for watermarking system, a 3D-mesh model is projected alternatively onto two constraints convex sets until the convergence condition is satisfied. The robustness convex set are designed for embedding the watermark into the distance distribution of the vertices to robust against the attacks, such as mesh simplification, cropping, rotation, translation, scaling, and vertex randomization. The invisibility convex set are designed for the embedded watermark to be invisible. The decision values and index that the watermark was embedded with are used to extract the watermark without the original model. Experimental results verify that the watermarked mesh model has invisibility and robustness against the attacks, such as translation, scaling, mesh simplification, cropping, and vertex randomization.

• Journal of Korea Game Society
• /
• v.7 no.4
• /
• pp.63-70
• /
• 2007

Since the fisheye lens has a wide field of view, it can capture the scene and illumination from all directions from far less number of omnidirectional images. Due to these advantages of the omnidirectional camera, it is widely used in surveillance and reconstruction of 3D structure of the scene In this paper, we present a new self-calibration algorithm of omnidirectional camera from uncalibrated images by considering the inlier distribution. First, one parametric non-linear projection model of omnidirectional camera is estimated with the known rotation and translation parameters. After deriving projection model, we can compute an essential matrix of the camera with unknown motions, and then determine the camera information: rotation and translations. The standard deviations are used as a quantitative measure to select a proper inlier set. The experimental results showed that we can achieve a precise estimation of the omnidirectional camera model and extrinsic parameters including rotation and translation.

• International Journal of Advanced Culture Technology
• /
• v.10 no.3
• /
• pp.104-110
• /
• 2022

This paper aims to analyze discourse metaphors by paying attention to Seoul mayoral by-election, mainly focusing on election campaign and its related news articles. The 2021 Seoul mayoral by-election was held because the former mayor died in an apparent suicide after he was accused of years of sexual harassment to a former secretary. But in the run-up to the by-election, the newly coined word 'alleged victim' from the ruling party caused a big controversy because the party attempted to deny the authenticity of the secretary's claim by calling her "an alleged victim," instead of "a victim" to defend the former mayor who is a member of the ruling party, implying that the woman's claim is just an allegation with no proof. Thus, this paper has analyzed how news stories were reported with regard to the word 'alleged victim' poser on news stories in two Korean quality newspapers, a conservative newspaper (Chosun Ilbo) and a liberal newspaper (Hankyoreh) from March 1 to April 1, 2021 and analyzed them with the framework of Lakoff and Johnson's Conceptual Metaphor Theory(1980). The findings are as follows: (i) the conservative newspaper reports this issue much more than the liberal newspaper; (ii) both quality newspapers follow the metaphor principles by Conceptual Metaphor Theory; (iii) the conservative newspaper is more likely to follow the Strick Father model (a conservative model) while the liberal newspaper is to follow the Nurturant Parent model (a liberal model), thus indicating that each newspaper's ideology is well represented by the models of Conceptual Metaphor Theory

• International Journal of Computer Science & Network Security
• /
• v.24 no.2
• /
• pp.101-112
• /
• 2024

Telugu language is considered as fourth most used language in India especially in the regions of Andhra Pradesh, Telangana, Karnataka etc. In international recognized countries also, Telugu is widely growing spoken language. This language comprises of different dependent and independent vowels, consonants and digits. In this aspect, the enhancement of Telugu Handwritten Character Recognition (HCR) has not been propagated. HCR is a neural network technique of converting a documented image to edited text one which can be used for many other applications. This reduces time and effort without starting over from the beginning every time. In this work, a Unicode based Handwritten Character Recognition(U-HCR) is developed for translating the handwritten Telugu characters into English language. With the use of Centre of Gravity (CG) in our model we can easily divide a compound character into individual character with the help of Unicode values. For training this model, we have used both online and offline Telugu character datasets. To extract the features in the scanned image we used convolutional neural network along with Machine Learning classifiers like Random Forest and Support Vector Machine. Stochastic Gradient Descent (SGD), Root Mean Square Propagation (RMS-P) and Adaptative Moment Estimation (ADAM)optimizers are used in this work to enhance the performance of U-HCR and to reduce the loss function value. This loss value reduction can be possible with optimizers by using CNN. In both online and offline datasets, proposed model showed promising results by maintaining the accuracies with 90.28% for SGD, 96.97% for RMS-P and 93.57% for ADAM respectively.