• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.11
• /
• pp.2251-2259
• /
• 2009

We propose an algorithm and system which generates 3D stereo image by composition of 2D image from 4 multiple clusters which 1 cluster was composed of 4 multiple cameras based on network. Proposed Schemes have a network-based client-server architecture for load balancing of system caused to process a large amounts of data with real-time as well as multiple cluster environments. In addition, we make use of JPEG compression and RAM disk method for better performance. Our scheme first converts input images from 4 channel, 16 cameras to binary image. And then we generate 3D stereo images after applying edge detection algorithm such as Sobel algorithm and Prewiit algorithm used to get disparities from images of 16 multiple cameras. With respect of performance results, the proposed scheme takes about 0.05 sec. to transfer image from client to server as well as 0.84 to generate 3D stereo images after composing 2D images from 16 multiple cameras. We finally confirm that our scheme is efficient to generate 3D stereo images in multiple view and multiple clusters environments with real-time.

• Geophysics and Geophysical Exploration
• /
• v.23 no.2
• /
• pp.97-114
• /
• 2020

Recently, many studies have been actively conducted on the application of machine learning in all branches of science and engineering. Studies applying machine learning are also rapidly increasing in all sectors of seismic exploration, including interpretation, processing, and acquisition. Among them, fault detection is a critical technology in seismic interpretation and also the most suitable area for applying machine learning. In this study, we introduced various machine learning techniques, described techniques suitable for fault detection, and discussed the reasons for their suitability. We collected papers published in renowned international journals and abstracts presented at international conferences, summarized the current status of the research by year and field, and intensively analyzed studies on fault detection using machine learning. Based on the type of input data and machine learning model, fault detection techniques were divided into seismic attribute-, image-, and raw data-based technologies; their pros and cons were also discussed.

• Journal of Korea Water Resources Association
• /
• v.45 no.7
• /
• pp.729-741
• /
• 2012

Among the methods of precipitation data acquisition, a rain gauge station has a distinctive advantage of direct measurement of rainfall itself, but multiple stations should be installed in order to obtain areal precipitation data required for hydrological analysis. On the other hand, a rainfall radar may provide areal distribution of rainfall in real time though it is an indirect measurement of radar echoes on rain drops. Rainfall radars have been shown useful especially for forecasting short-term localized torrential storms that may cause catastrophic flash floods. CAPPI (Constant Altitude Plan Position Indicator), which is one of the several types of radar rainfall image data, has been provided on the Internet in real time by Korea Meteorological Administration (KMA). It is one of the most widely available rainfall data in Korea with fairly high level of confidence as it is produced with bias adjustment and quality control procedures by KMA. The objective of this study is to develop an improved way to extract quantitative rainfall data applicable to even very small watersheds from CAPPI using CIVCOM, which is a new image processing method based on a vector-based scheme proposed in this study rather than raster-based schemes proposed by other researchers. This study shows usefulness of CIVCOM through comparison of rainfall data produced by image processing methods including traditional raster-based schemes and a newly proposed vector-based one.

• Journal of KIISE:Databases
• /
• v.31 no.6
• /
• pp.684-698
• /
• 2004

The Hippocratic database model recently proposed by Agrawal et al. incorporates privacy protection capabilities into relational databases. Since the Hippocratic database is based on the relational database, it needs extensions to be adapted for XML databases. In this paper, we propose the Hippocratic XML database model, an extension of the Hippocratic database model for XML databases and present an efficient access control mechanism under this model. In contrast to relational data, XML data have tree-like hierarchies. Thus, in order to manage these hierarchies of XML data, we extend and formally define such concepts presented in the Hippocratic database model as privacy preferences, privacy policies, privacy authorizations, and usage purposes of data records. Next, we present a new mechanism, which we call the authorization index, that is used in the access control mechanism. This authorization index, which is Implemented using a multi-dimensional index, allows us to efficiently search authorizations implied by the authorization granted on the nearest ancestor using the nearest neighbor search technique. Using synthetic and real data, we have performed extensive experiments comparing query processing time with those of existing access control mechanisms. The results show that the proposed access control mechanism improves the wall clock time by up to 13.6 times over the top-down access control strategy and by up to 20.3 times over the bottom-up access control strategy The major contributions of our paper are 1) extending the Hippocratic database model into the Hippocratic XML database model and 2) proposing an efficient across control mechanism that uses the authorization index and nearest neighbor search technique under this model.

• Journal of Broadcast Engineering
• /
• v.27 no.1
• /
• pp.3-12
• /
• 2022

Image deblurring aims to remove image blur, which can be generated while shooting the pictures by the movement of objects, camera shake, blurring of focus, and so forth. With the rise in popularity of smartphones, it is common to carry portable digital cameras daily, so image deblurring techniques have become more significant recently. Originally, image deblurring techniques have been studied using traditional optimization techniques. Then with the recent attention on deep learning, deblurring methods based on convolutional neural networks have been actively proposed. However, most of them have been developed while focusing on better performance. Therefore, it is not easy to use in real situations due to the speed of their algorithms. To tackle this problem, we propose a novel deep learning-based deblurring algorithm that can be operated in real-time on HD resolution. In addition, we improved the training and inference process and could increase the performance of our model without any significant effect on the speed and the speed without any significant effect on the performance. As a result, our algorithm achieves real-time performance by processing 33.74 frames per second at 1280×720 resolution. Furthermore, it shows excellent performance compared to its speed with a PSNR of 29.78 and SSIM of 0.9287 with the GoPro dataset.

• Microbiology and Biotechnology Letters
• /
• v.31 no.4
• /
• pp.377-382
• /
• 2003

Purification of methioninase resulted in a yield of 69%, and SDS-PAGE analysis of the purified product revealed a single band of approximately 43 kDa in molecular weight. in vitro experiments with cancer cells incubated in methionine-free media demonstrated an increase in $^{11}$ C-methionine uptake to 25.8$\pm$1.1% at 6 hr, 31.8$\pm$0.8% at 24 hr, and 62.2$\pm$0.6% at 48hr, compared to controls. Treatment of the cancer cells with purified methioninase showed no decrease in survival after a 2 hr incubation with 0.01 U/ml, but survival of RR1022 cells decreased 30% after 24 to 48 hr incubation. SKOV-3 cells showed a 5% and 14% decrease in survival with 0.1 and 1 U/ml methioninase after 24 hr. After 48hr survival decreased 15% and 24% with 0.1 and 1 U/ml methioninase. Measurements of $^{11}$ C-methionine uptake in RR1022 cells demonstrated no change at 2 hr, but a 13.7$\pm$4.7% and 40.7$\pm$2.6% increase in uptake at 24 and 48 hr, respectively. SKOV-3 cells also showed no change at 2 hr, but had a 17.7$\pm$7.2% and 38.9$\pm$4.9% increase in $^{11}$ C-methionine uptake after 24 hr and 48 hr treatment with methioninase, respectively. $^{11}$ C-methionine PET imaging revealed clear visualization of both the tumors and contralateral infectious lesions. Administration of rMET appeared to result in a slight increase in tumor:nontumor contrast on $^{11}$ C-methionine PET images. Injection of purified methioninase also produced PET images where tumor uptake was higher than that of infectious lesions.

• Geophysics and Geophysical Exploration
• /
• v.26 no.2
• /
• pp.37-51
• /
• 2023

The ground roll is the most common coherent noise in land seismic data and has an amplitude much larger than the reflection event we usually want to obtain. Therefore, ground roll suppression is a crucial step in seismic data processing. Several techniques, such as f-k filtering and curvelet transform, have been developed to suppress the ground roll. However, the existing methods still require improvements in suppression performance and efficiency. Various studies on the suppression of ground roll in seismic data have recently been conducted using deep learning methods developed for image processing. In this paper, we introduce three models (DnCNN (De-noiseCNN), pix2pix, and CycleGAN), based on convolutional neural network (CNN) or conditional generative adversarial network (cGAN), for ground roll suppression and explain them in detail through numerical examples. Common shot gathers from the same field were divided into training and test datasets to compare the algorithms. We trained the models using the training data and evaluated their performances using the test data. When training these models with field data, ground roll removed data are required; therefore, the ground roll is suppressed by f-k filtering and used as the ground-truth data. To evaluate the performance of the deep learning models and compare the training results, we utilized quantitative indicators such as the correlation coefficient and structural similarity index measure (SSIM) based on the similarity to the ground-truth data. The DnCNN model exhibited the best performance, and we confirmed that other models could also be applied to suppress the ground roll.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.7
• /
• pp.326-334
• /
• 2024

Extra-oral imaging techniques such as Panoramic X-rays (PXs) and Cone Beam Computed Tomography (CBCT) are the most preferred imaging modalities in dental clinics owing to its patient convenience during imaging as well as their ability to visualize entire teeth information. PXs are preferred for routine clinical treatments and CBCTs for complex surgeries and implant treatments. However, PXs are limited by the lack of third dimensional spatial information whereas CBCTs inflict high radiation exposure to patient. When a PX is already available, it is beneficial to reconstruct the 3D oral structure from the PX to avoid further expenses and radiation dose. In this paper, we propose 3DentAI - an U-Net based deep learning framework for 3D reconstruction of oral structure from a PX image. Our framework consists of three module - a reconstruction module based on attention U-Net for estimating depth from a PX image, a realignment module for aligning the predicted flattened volume to the shape of jaw using a predefined focal trough and ray data, and lastly a refinement module based on 3D U-Net for interpolating the missing information to obtain a smooth representation of oral cavity. Synthetic PXs obtained from CBCT by ray tracing and rendering were used to train the networks without the need of paired PX and CBCT datasets. Our method, trained and tested on a diverse datasets of 600 patients, achieved superior performance to GAN-based models even with low computational complexity.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.183-203
• /
• 2018

News articles are the most suitable medium for examining the events occurring at home and abroad. Especially, as the development of information and communication technology has brought various kinds of online news media, the news about the events occurring in society has increased greatly. So automatically summarizing key events from massive amounts of news data will help users to look at many of the events at a glance. In addition, if we build and provide an event network based on the relevance of events, it will be able to greatly help the reader in understanding the current events. In this study, we propose a method for extracting event networks from large news text data. To this end, we first collected Korean political and social articles from March 2016 to March 2017, and integrated the synonyms by leaving only meaningful words through preprocessing using NPMI and Word2Vec. Latent Dirichlet allocation (LDA) topic modeling was used to calculate the subject distribution by date and to find the peak of the subject distribution and to detect the event. A total of 32 topics were extracted from the topic modeling, and the point of occurrence of the event was deduced by looking at the point at which each subject distribution surged. As a result, a total of 85 events were detected, but the final 16 events were filtered and presented using the Gaussian smoothing technique. We also calculated the relevance score between events detected to construct the event network. Using the cosine coefficient between the co-occurred events, we calculated the relevance between the events and connected the events to construct the event network. Finally, we set up the event network by setting each event to each vertex and the relevance score between events to the vertices connecting the vertices. The event network constructed in our methods helped us to sort out major events in the political and social fields in Korea that occurred in the last one year in chronological order and at the same time identify which events are related to certain events. Our approach differs from existing event detection methods in that LDA topic modeling makes it possible to easily analyze large amounts of data and to identify the relevance of events that were difficult to detect in existing event detection. We applied various text mining techniques and Word2vec technique in the text preprocessing to improve the accuracy of the extraction of proper nouns and synthetic nouns, which have been difficult in analyzing existing Korean texts, can be found. In this study, the detection and network configuration techniques of the event have the following advantages in practical application. First, LDA topic modeling, which is unsupervised learning, can easily analyze subject and topic words and distribution from huge amount of data. Also, by using the date information of the collected news articles, it is possible to express the distribution by topic in a time series. Second, we can find out the connection of events in the form of present and summarized form by calculating relevance score and constructing event network by using simultaneous occurrence of topics that are difficult to grasp in existing event detection. It can be seen from the fact that the inter-event relevance-based event network proposed in this study was actually constructed in order of occurrence time. It is also possible to identify what happened as a starting point for a series of events through the event network. The limitation of this study is that the characteristics of LDA topic modeling have different results according to the initial parameters and the number of subjects, and the subject and event name of the analysis result should be given by the subjective judgment of the researcher. Also, since each topic is assumed to be exclusive and independent, it does not take into account the relevance between themes. Subsequent studies need to calculate the relevance between events that are not covered in this study or those that belong to the same subject.

• Journal of KIISE:Computer Systems and Theory
• /
• v.27 no.7
• /
• pp.650-664
• /
• 2000

NOW(Network of Workstations) is considered as a platform for running parallel programs by many people. One of the fundamental problems that must be addressed to achieve good performance for parallel programs on NOW is the determination of efficient job scheduling policies. Currently most research on NOW assumes that all the workstations in the NOW have the same processing power. In this paper we consider a NOW in which workstations may have different computing power. We introduce 10 classes of space sharing-based scheduling policies that can be applied to the NOW with heterogeneous computing power. We compare the performance of these scheduling policies by using the simulator which accepts synthetically generated sequential and parallel workloads and generates the response time and waiting time of parallel jobs as performance indices of various scheduling strategies. Through the experiments the case when a parallel program is partitioned heterogeneously in proportion to the computing power of workstations is shown to have better performance than when a parallel program is partitioned into parallel processes of the same size. When the owner returns to the workstation which is executing a parallel process, the policy which just lowers the priority of the parallel process shows better performance than the one which migrates the parallel process to a new idle workstation. Among the policies which use heterogeneous partitioning and process priority lowering, the adaptive policy performed best across the wide range of inter-arrival time of parallel programs but when the load imbalance among parallel processes becomes very high, the modified adaptive policy performed better.