• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.10
• /
• pp.3668-3684
• /
• 2021

Video action recognition is widely used in video surveillance, behavior detection, human-computer interaction, medically assisted diagnosis and motion analysis. However, video action recognition can be disturbed by many factors, such as background, illumination and so on. Two-stream convolutional neural network uses the video spatial and temporal models to train separately, and performs fusion at the output end. The multi segment Two-Stream convolutional neural network model trains temporal and spatial information from the video to extract their feature and fuse them, then determine the category of video action. Google Xception model and the transfer learning is adopted in this paper, and the Xception model which trained on ImageNet is used as the initial weight. It greatly overcomes the problem of model underfitting caused by insufficient video behavior dataset, and it can effectively reduce the influence of various factors in the video. This way also greatly improves the accuracy and reduces the training time. What's more, to make up for the shortage of dataset, the kinetics400 dataset was used for pre-training, which greatly improved the accuracy of the model. In this applied research, through continuous efforts, the expected goal is basically achieved, and according to the study and research, the design of the original dual-flow model is improved.

• Journal of Korean Association for Spatial Structures
• /
• v.21 no.1
• /
• pp.79-86
• /
• 2021

Control performance of a smart tuned mass damper (TMD) mainly depends on control algorithms. A lot of control strategies have been proposed for semi-active control devices. Recently, machine learning begins to be applied to development of vibration control algorithm. In this study, a reinforcement learning among machine learning techniques was employed to develop a semi-active control algorithm for a smart TMD. The smart TMD was composed of magnetorheological damper in this study. For this purpose, an 11-story building structure with a smart TMD was selected to construct a reinforcement learning environment. A time history analysis of the example structure subject to earthquake excitation was conducted in the reinforcement learning procedure. Deep Q-network (DQN) among various reinforcement learning algorithms was used to make a learning agent. The command voltage sent to the MR damper is determined by the action produced by the DQN. Parametric studies on hyper-parameters of DQN were performed by numerical simulations. After appropriate training iteration of the DQN model with proper hyper-parameters, the DQN model for control of seismic responses of the example structure with smart TMD was developed. The developed DQN model can effectively control smart TMD to reduce seismic responses of the example structure.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.9
• /
• pp.3152-3171
• /
• 2022

This work investigates the physical layer secrecy of a multi-antenna hybrid satellite-terrestrial relay networks (HSTRN) with jamming, in which a satellite aims to make communication with a destination user by means of a relay, along with spatially random eavesdroppers. In order to weaken the signals of eavesdroppers, the conventional relay can also generate intentional interference, besides forwarding the received signal. Shadowed-Rician fading is adopted in satellite link, while Rayleigh fading is adopted in terrestrial link, eavesdropper link and jamming link. The analytical and asymptotic formulas for the system secrecy outage probability (SOP) are characterized. Practical insights on the diversity order of the network are revealed according to the asymptotic behavior of SOP at high signal-to-noise ratio (SNR) regime. Then, analysis of the system throughput is examined to assess the secrecy performance. In the end, numerical simulation results are presented to validate the theoretical analysis and point out: (1) The secrecy performance of the considered network is affected by the channel fading scenario, the system configuration; (2) Decrease of the relay coverage airspace can provide better SOP performance; (3) Jamming from the relay can improve secrecy performance without additional network resources.

• Communications for Statistical Applications and Methods
• /
• v.30 no.6
• /
• pp.561-575
• /
• 2023

Brain network analysis has attracted the interest of neuroscience researchers in studying brain diseases. Magnetoencephalography (MEG) is especially proper for analyzing functional connectivity due to high temporal and spatial resolution. The application of graph theory for functional connectivity analysis has been studied widely, but research on network modeling for MEG still needs more. Temporal exponential random graph model (TERGM) considers temporal dependencies of networks. We performed the brain network analysis, including static/temporal network statistics, on two groups of epilepsy patients who removed the left (LT) or right (RT) part of the brain and healthy controls. We investigate network differences using Multiset canonical correlation analysis (MCCA) and TERGM between epilepsy patients and healthy controls (HC). The brain network of healthy controls had fewer temporal changes than patient groups. As a result of TERGM, on the simulation networks, LT and RT had less stable state than HC in the network connectivity structure. HC had a stable state of the brain network.

• Journal of the Economic Geographical Society of Korea
• /
• v.6 no.1
• /
• pp.99-117
• /
• 2003

Major proposes of this study are to analyze Korean Shipbuilding's production network and cooperation between related firms and to understand their spatial features. The Geographical study about networks has focused on automobiles, electronics or communications industries. That's because those industries have distinctive spatial features thanks to the full growth of subcontract structures. The labor-centered, capital-centered shipbuilding industry, differently from other manufacturing industries, has production networks where outside trades are common. Today internet based communication is being reinforced and the flexibility of purchase circumstances is being positively proceeded. The central axes of the production network of Sumsung shipbuilding are internationally Europe and Asian area, nationally Busan and Gyeongnam province and Seoul Metropolitan areas, locally inside subcontracting firms. And In order to construct mutual trust with cooperation firms, Sumsung shipbuilding is trying to reinforce two-way cooperation relation through 'SungJoHoi' organized with outside subcontracting firms and 'inside cooperation firms conference'. In conclusion, Sumsung shipbuilding is trying to strengthen the competitiveness of shipbuilding industry, spatially by setting up the network through globalization-localization and functionally by constructing the partnership production network for mutual communication.

• KIEAE Journal
• /
• v.16 no.6
• /
• pp.69-75
• /
• 2016

Climate change has resulted in the urban heat island (UHI) effect throughout the globe, contributing to heat-related illness and fatalities. In order to reduce such damage, it is necessary to improve the climate observation network for precise observation of the urban thermal environment and quick UHI forecasting system. Purpose: This study analyzed the effective range of the climate observation network and the distribution of the existing Automatic Weather Stations (AWS) in Seoul to propose optimal locations for additional installment of AWS. Method: First, we performed quality analysis to pinpoint missing values and outliers within the high-density temperature data measured. With the result from the analysis, a spatial autocorrelation structure in the temperature data was tested to draw the effective range and correlation distance for each major time period. Result: As a result, it turned out that the optimal effective range for the climate observation network in Seoul in July was a radius of 2.8 kilometers. Based on this result, population density, and temperature data, we selected the locations for additional installment of AWS. This study is expected to be used to generate urban temperature maps, select and move measurement locations since it is able to suggest valid, specific spatial ranges when the data measured in point is converted into surface data.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.1
• /
• pp.119-124
• /
• 2010

The MAC layer in ad-hoc network which makes network of nodes without infrastructure for a time has became an issue to reduce delay, allocate fairly bandwidth, control TX/RX power and improve throughput. Specially, the problem to reduce power consumption in ad-hoc network is very important part as ad-hoc devices use the limited battery. For solution of the problem, many power control algorithms, such as distribute power control, PCM (Power Control MAC) and F-PCF (Fragmentation based PCM), are proposed to limit power consumption until now. Although the algorithms are designed to minimize power consumption, the latency communication zone is generated by power control of RX/TX nodes. However the algorithms don't suitably reuse the space. In this paper proposes the algorithm to improve data throughput through Spatial Reuse based on a power control method.

• Journal of the Economic Geographical Society of Korea
• /
• v.8 no.2
• /
• pp.195-215
• /
• 2005

This study was carried into execution to analyze Spatial Characteristics of Seoul Business Services Cluster in circumstance of local economic development led by knowledge-intensive business services. The analysis was performed for industrial agglomeration and companies' linkage and network. As the result, there are three business services cluster, including the largest one in Gang-Nam Gu, Seoul, and for last 10 years, there has been increasing development of business services cluster. In the meanwhile, their linkage and network have not been performed briskly. Therefore, from now on, a plan that can help companies' linkage and network performed inside of cluster to have active and international structure has to be considered in cluster policy.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.6
• /
• pp.4228-4236
• /
• 2015

Due to the rapid advance in web and mobile computing technologies, normal users have become to produce, provide, and share a varied form of spatial data and information. In the domain of spatial information, numerous researches on GIS have been conducted to provide spatial information services based on a geo-sensor network and a data integration and processing technology. However, to provide user-oriented information, a context information model is necessary to associate GIS data with web and sensor data. Context awareness services is designed to provide specific information, minimizing users' interference. For which, the context information model expresses the relationship of various data from sensor networks and mobile applications and provides a user-specific information considering location and area of interest. Thus, this research aims to develops a context information model based air-pollution information system that obtains and analyses air pollution data and reflects the analysis results on an air-pollution policy. Also, this system aims to raise citizens' awareness on air-pollution and to promote citizens' participatory to improve city's air quality.

• Journal of Cadastre & Land InformatiX
• /
• v.48 no.2
• /
• pp.213-225
• /
• 2018

For constructing three-dimensional (3D) spatial information occlusion region problem arises in the process of taking the texture of the building. In order to solve this problem, it is necessary to investigate the automation method to automatically recognize the occlusion region, issue it, and automatically complement the texture. In fact there are occasions when it is possible to generate a very large number of structures and occlusion, so alternatives to overcome are being considered. In this study, we attempt to apply an approach to automatically create an occlusion region based on learning by patterning the blocked region using the recently emerging deep learning algorithm. Experiment to see the performance automatic detection of people, banners, vehicles, and traffic lights that cause occlusion in building walls using two advanced algorithms of Convolutional Neural Network (CNN) technique, Faster Region-based Convolutional Neural Network (R-CNN) and Mask R-CNN. And the results of the automatic detection by learning the banners in the pre-learned model of the Mask R-CNN method were found to be excellent.