• Journal of Korea Water Resources Association
• /
• v.43 no.4
• /
• pp.383-391
• /
• 2010

The groundwater recharge was assessed by using both SWAT and HELP models in Bocheong-cheon watershed. The SWAT model is a comprehensive surface and subsurface model, but it lacks the physical basis for simulating a soil water percolation process. The HELP model which has a drawback in simulating subsurface lateral flow and groundwater flow component can simulate soil water percolation process by considering the unsaturated flow effect of soil layers. The SWAT model has been successfully applied for estimating groundwater recharge in a number of watersheds in Korea, while the application of HELP model has been very limited. The subsurface lateral flow parameter was proposed in order to consider the subsurface lateral flow effect in HELP model and the groundwater recharge was simulated by the modified exponential decay weighting function in HELP model. The simulation results indicate that the recharge of HELP model significantly depends on the values of lateral flow parameter. The recharge errors between SWAT and HELP are the smallest when the lateral flow parameter is about 0.6 and the recharge rates between two models are shown to be reasonably comparable for daily, monthly, and yearly time scales. The HELP model is useful for estimating groundwater recharge at watershed scale because the model structure and input parameters of HELP model are simpler than that of SWAT model. The accuracy of assessing the groundwater recharge might be improved by the concurrent application of SWAT model and HELP model.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.8
• /
• pp.1809-1818
• /
• 2010

In order to detect the skin color area from input images, many prior researches have divided an image into the pixels having a skin color and the other pixels. In a still image or videos, it is very difficult to exactly extract the skin pixels because lighting condition and makeup generate a various variations of skin color. In this thesis, we propose a method that improves its performance using hierarchical merging of 3D skin color model and context informations for the images having various difficulties. We first make 3D color histogram distributions using skin color pixels from many YCbCr color images and then divide the color space into 3 layers including skin color region(Skin), non-skin color region(Non-skin), skin color candidate region (Skinness). When we segment the skin color region from an image, skin color pixel and non-skin color pixels are determined to skin region and non-skin region respectively. If a pixel is belong to Skinness color region, the pixels are divided into skin region or non-skin region according to the context information of its neighbors. Our proposed method can help to efficiently segment the skin color regions from images having many distorted skin colors and similar skin colors.

• The Sea:JOURNAL OF THE KOREAN SOCIETY OF OCEANOGRAPHY
• /
• v.8 no.2
• /
• pp.187-198
• /
• 2003

High-resolution (Chirp, 3-11 kHz) echo facies and sedimentary facies of piston-core sediments were analyzed to reveal the late Quaternary depositional processes in the Korea Plateau and Ulleung Interplain Gap. The Korea Plateau is an Isolated topographic high with a very restricted input of terrigenous sediments, and its slope is characterized by a thin sediment cover and various-scale submarine canyons and valleys. Echo and sedimentary facies suggest that the plateau has been moulded mainly by persistent (hemi) pelagic sedimentation and intermittent settling of volcanic ashes. Sediments on the plateau slope and steep margins of ridges and seamounts were reworked by earthquake-induced, large-scale slope failures accompanied by slides, slumps and debris flows. As major fraction of the reworked sediments consists of (hemi) pelagic clay particles, large amounts of sediments released from mass flows were easily suspended to form turbid nepheloid layers rather than bottom-hugging turbidity currents, which flowed further downslope through the submarine canyons and spreaded over the Ulleung Basin plain. In the Ulleung Interplain Gap, sediments were introduced mainly by (hemi) pelagic settling and subordinate episodic mass flows (turbidity currents and debris flows) along the submarine channels from the slopes of the Oki Bank and Dok Island. The sediments in the Ulleung Interplain Channel and its margin were actively eroded and reworked by the deep water flow from the Japan Basin.

• Journal of Broadcast Engineering
• /
• v.24 no.3
• /
• pp.472-484
• /
• 2019

This paper proposes a time-domain sound event detection algorithm using DNN (Deep Neural Network). In this system, time domain sound waveform data which is not converted into the frequency domain is used as input to the DNN. The overall structure uses CRNN structure, and GLU, ResNet, and Squeeze-and-excitation blocks are applied. And proposed structure uses structure that considers features extracted from several layers together. In addition, under the assumption that it is practically difficult to obtain training data with strong labels, this study conducted training using a small number of weakly labeled training data and a large number of unlabeled training data. To efficiently use a small number of training data, the training data applied data augmentation methods such as time stretching, pitch change, DRC (dynamic range compression), and block mixing. Unlabeled data was supplemented with insufficient training data by attaching a pseudo-label. In the case of using the neural network and the data augmentation method proposed in this paper, the sound event detection performance is improved by about 6 %(based on the f-score), compared with the case where the neural network of the CRNN structure is used by training in the conventional method.

• Journal of the Korean Society of Radiology
• /
• v.14 no.7
• /
• pp.991-1001
• /
• 2020

Supervised deep learning technologies for improving the image quality of computed tomography (CT) need a lot of training data. When input images have different characteristics with training images, the technologies cause structural distortion in output images. In this study, an imaging model based on the deep reinforcement learning (DRL) was developed for overcoming the drawbacks of the supervised deep learning technologies and reducing noise in CT images. The DRL model was consisted of shared, value and policy networks, and the networks included convolutional layers, rectified linear unit (ReLU), dilation factors and gate rotation unit (GRU) in order to extract noise features from CT images and improve the performance of the DRL model. Also, the quality of the CT images obtained by using the DRL model was compared to that obtained by using the supervised deep learning model. The results showed that the image accuracy for the DRL model was higher than that for the supervised deep learning model, and the image noise for the DRL model was smaller than that for the supervised deep learning model. Also, the DRL model reduced the noise of the CT images, which had different characteristics with training images. Therefore, the DRL model is able to reduce image noise as well as maintain the structural information of CT images.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.23 no.1
• /
• pp.13-24
• /
• 2021

Tunnel boring machine (TBM) is widely used for tunnel excavation in hard rock and soft ground. In the perspective of TBM-based tunneling, one of the main challenges is to drive the machine optimally according to varying geological conditions, which could significantly lead to saving highly expensive costs by reducing the total operation time. Generally, drilling investigations are conducted to survey the geological ground before the TBM tunneling. However, it is difficult to provide the precise ground information over the whole tunnel path to operators because it acquires insufficient samples around the path sparsely and irregularly. To overcome this issue, in this study, we proposed a geological type classification system using the TBM operating data recorded in a 5 s sampling rate. We first categorized the various geological conditions (here, we limit to granite) as three geological types (i.e., rock, soil, and mixed type). Then, we applied the preprocessing methods including outlier rejection, normalization, and extracting input features, etc. We adopted a deep neural network (DNN), which has 6 hidden layers, to classify the geological types based on TBM operating data. We evaluated the classification system using the 10-fold cross-validation. Average classification accuracy presents the 75.4% (here, the total number of data were 388,639 samples). Our experimental results still need to improve accuracy but show that geology information classification technique based on TBM operating data could be utilized in the real environment to complement the sparse ground information.

• Journal of the Korean Society of Radiology
• /
• v.16 no.3
• /
• pp.295-302
• /
• 2022

Sparse-view computed tomography (CT) imaging technique is able to reduce radiation dose, ensure the uniformity of image characteristics among projections and suppress noise. However, the reconstructed images obtained by the sparse-view CT imaging technique suffer from severe artifacts, resulting in the distortion of image quality and internal structures. In this study, we proposed a convolutional neural network (CNN) with wavelet transformation and residual learning for reducing artifacts in sparse-view CT image, and the performance of the trained model was quantitatively analyzed. The CNN consisted of wavelet transformation, convolutional and inverse wavelet transformation layers, and input and output images were configured as sparse-view CT images and residual images, respectively. For training the CNN, the loss function was calculated by using mean squared error (MSE), and the Adam function was used as an optimizer. Result images were obtained by subtracting the residual images, which were predicted by the trained model, from sparse-view CT images. The quantitative accuracy of the result images were measured in terms of peak signal-to-noise ratio (PSNR) and structural similarity (SSIM). The results showed that the trained model is able to improve the spatial resolution of the result images as well as reduce artifacts in sparse-view CT images effectively. Also, the trained model increased the PSNR and SSIM by 8.18% and 19.71% in comparison to the imaging model trained without wavelet transformation and residual learning, respectively. Therefore, the imaging model proposed in this study can restore the image quality of sparse-view CT image by reducing artifacts, improving spatial resolution and quantitative accuracy.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.39 no.6
• /
• pp.515-523
• /
• 2021

An integrated underground spatial map consists of underground facilities, underground structures, and ground information, and is periodically updated. In this paper, we design and implement a system for detecting and extracting only changed ground objects to shorten the map update speed. To find the changed objects, all the objects are compared, which are included in the newly input map and the reference map in the integrated map. Since the entire process of comparing objects and generating results is classified by function, the implemented system is composed of several modules such as object comparer, changed object detector, history data manager, changed object extractor, changed type classifier, and changed object saver. We use two metrics: detection rate and extraction rate, to evaluate the performance of the system. As a result of applying the system to boreholes, ground wells, soil layers, and rock floors in Pyeongtaek, 100% of inserted, deleted, and updated objects in each layer are detected. In addition, it provides the advantage of ensuring the up-to-dateness of the reference map by downloading it whenever maps are compared. In the future, additional research is needed to confirm the stability and effectiveness of the developed system using various data to apply it to the field.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.10
• /
• pp.373-380
• /
• 2022

This paper presents a deep learning based group synchronization that supports networked immersive interactions between remote users. The goal of group synchronization is to enable all participants to synchronously interact with others for increasing user presence Most previous methods focus on NTP-based clock synchronization to enhance time accuracy. Moving average filters are used to control media playout time on the synchronization server. As an example, the exponentially weighted moving average(EWMA) would be able to track and estimate accurate playout time if the changes in input data are not significant. However it needs more time to be stable for any given change over time due to codec and system loads or fluctuations in network status. To tackle this problem, this work proposes the Deep Group Synchronization(DeepGroupSync), a group synchronization based on deep learning that models important features from the data. This model consists of two Gated Recurrent Unit(GRU) layers and one fully-connected layer, which predicts an optimal playout time by utilizing the sequential playout delays. The experiments are conducted with an existing method that uses the EWMA and the proposed method that uses the DeepGroupSync. The results show that the proposed method are more robust against unpredictable or rapid network condition changes than the existing method.

• Journal of the Korean Geotechnical Society
• /
• v.38 no.12
• /
• pp.45-65
• /
• 2022

In this paper, a diaphragm wall that supports soils and rock was modeled using FLAC, a finite difference analysis program, to evaluate the seismic behavior of temporary retaining structures in a deep excavation. The appropriateness of the numerical model was verified by comparing its results with those of the centrifuge test performed in a similar condition. The bending moment distribution along the diaphragm wall shows a very similar tendency, and the maximum acceleration obtained at the backfill and top of the wall shows a difference within 5%. Based on the developed model, a parametric study was conducted in various input earthquake, ground, and excavation conditions. The maximum structural forces and bending moment under earthquake loading were compared with the maximum values during excavation, from which the critical condition that requires a seismic design was roughly sorted out. The maximum bending moment of a wall that retains soil layers increased 17%. Particularly, the axial force of struts located in loose soils increased 32% under 100 years return period of an earthquake event, which strongly is estimated to require seismic design for structural safety.