• Journal of the Korea Society of Computer and Information
• /
• v.25 no.1
• /
• pp.55-61
• /
• 2020

With the development of deep learning technology and advances in computing power, video-based research is now gaining more and more attention. Video data contains a large amount of temporal and spatial information, which is the biggest difference compared with image data. It has a larger amount of data. It has attracted intense attention in computer vision. Among them, motion recognition is one of the research focuses. However, the action recognition of human in the video is extremely complex and challenging subject. Based on many research in human beings, we have found that artificial intelligence-like attention mechanisms are an efficient model for cognition. This efficient model is ideal for processing image information and complex continuous video information. We introduce this attention mechanism into video action recognition, paying attention to human actions in video and effectively improving recognition efficiency. In this paper, we propose a new 3D residual attention network using convolutional neural network based on two attention models to identify human action behavior in the video. An evaluation result of our model showed up to 90.7% accuracy.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.8 no.12
• /
• pp.1841-1848
• /
• 2013

In this paper, the plans for improvement of real-time security monitoring accuracy and expansion of control region were investigated through comprehensive and systematic collection and analysis of the anomalous activities that inflow and outflow in the network on a large scale in order to overcome the existing security monitoring system based on stylized detection patterns which could correspond to only very limited cyber attacks. This study established an anomaly observation system to collect, store and analyze a diverse infringement threat information flowing into the darknet network, and presented the information classification system of cyber threats, unknown anomalies and high-risk anomalous activities through the statistics based trend analysis of hacking. If this security monitoring system utilizing darknet traffic as presented in the study is applied, it was indicated that detection of all infringement threats was increased by 12.6 percent compared with conventional case and 120 kinds of new type and varietal attacks that could not be detected in the past were detected.

• Computational Structural Engineering
• /
• v.3 no.3
• /
• pp.113-123
• /
• 1990

The paper is concerned with the elasto-plastic and geometrically nonlinear analysis of shell structures using an improved degenerated shell element. In the formulation of the element stiffness, the combined use of three different techniques was made. They are; 1) an enhanced interpolation of transverse shear strains in the natural coordinate system to overcome the shear locking problem ; 2) the reduced integration technique in in-plane strains to avoid the membrane locking behavior ; and 3) selective addition of the nonconforming displacement modes to improve the element performances. This element is free of serious shear/membrane locking problems and undesirable compatible/commutable spurious kinematic deformation modes. In the formulation for plastic deformation, the concept of a layered element model is used and the material is assumed von Mises yield criterion. An incremental total Lagrangian formulation is presented which allows the calculation of arbitrarily large displacements and rotations. The resulting non-linear equilibrium equations are solved by the Netwon-Raphson method combined with load or displacement increment. The versatility and accuracy of this improved degenerated shell element are demonstrated by solving several numerical examples.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.41 no.5
• /
• pp.321-328
• /
• 2017

In this study, vortex tube system model is developed to predict the temperature of the hot and the cold sides. The vortex tube model is developed based on the system identification method, and the model utilized in this work to design the vortex tube is ARX type (Auto-Regressive with eXtra inputs). The derived polynomial model is validated against experimental data to verify the overall model accuracy. It is also shown that the derived model passes the stability test. It is confirmed that the derived model closely mimics the physical behavior of the vortex tube from both the static and dynamic numerical experiments by changing the angles of the low-temperature side throttle valve, clearly showing temperature separation. These results imply that the system identification based modeling can be a promising approach for the prediction of complex physical systems, including the vortex tube.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.10 no.1
• /
• pp.23-28
• /
• 2010

Reliable dynamic analysis is essential in order to properly maintain structures so that structural hazards may be minimized. The finite element method (FEM) is proven to be an affective approximate method of structural analysis if proper element types and meshes are chosen. When the method is applied to dynamics analyzed in time domain, the meshes may need to be modified at each time step. As many meshes need to be generated, adaptive mesh generation schemes have become an important part in complex time domain dynamic finite element analyses of structures. In this paper, an adaptive mesh generation scheme for dynamic finite element analyses of structures is described. The concept of representative strain value is used for error estimates and the refinements of meshes use combinations of the h-method (node movement) and the r-method (element division). The validity of the scheme is shown through a cantilever beam example under a concentrated load with varying values. The example shows reasonable accuracy and efficient computing time. Furthermore, the study shows the potential for the scheme's effective use in complex structural dynamic problems such as those under seismic or erratic wind loads.

• Nuclear Engineering and Technology
• /
• v.24 no.3
• /
• pp.243-251
• /
• 1992

Behavior of loop seal formation and clearing during small break loss of coolant accident is investigated using the RELAP5/MOD 2 and /MOD3 codes with the test of SB-CL-18 of the LSIF (Large Scale Test Facility）. The present study examines the thermal-hydraulic mechanisms responsible for early core uncovery including the manometric effect due to an asymmetric coolant holdup in the steam generator upflow and downflow side. The analysis with the RELAP5/MOD2 demonstrates the main phenomena occuring in the depressurization transient including the loop seal formation and clearing with sufficient accuracy. Nevertheless, several differences regarding the evolution of phenomena and their timing have been pointed out in かe base calculations. The RELAP5/MOD3 predicts overall phenomena, particularly the steam generator liquid holdup better than the RELAP5/MOD2. The nodalization study in the components of the steam generator U-tubes and the cross-over legs wiか the RELAP5/MOD3 results in good prediction of the loop seal clearing phenomena and their timing.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.8
• /
• pp.5654-5663
• /
• 2015

In the unsaturated soil, suction, the negative pore water pressure leads to increases of the yield stress and the plastic shear stiffness of the soil skeleton due to the growth in interparticle stress. Therefore, in this study, the stress-strain relationship based on cyclic elasto-plastic constitutive model extended for unsaturated soil considering the 1st and the 2nd yield functions was induced in order to account for these effects of suction under the dynamic loading condition such as the earthquake. Through the program code considering this relationship and the routine of the cyclic loading with the reversion of loading direction, the numerical simulation of the cyclic triaxial test under the unsaturated condition would be possible. It is expected that the results of this study possibly contribute to the accuracy improvement on the prediction of unsaturated soil behavior under the dynamic loading condition.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.15 no.4
• /
• pp.693-702
• /
• 2002

The aim of this study is to develop a 3-dimensional dynamic analysis model, capable of considering the interaction between vehicles and bridges more accurately. The dynamic analysis model is developed with the high-speed train (KTX) and a 2-span continuous prestressed concrete box girder bridge with a double track. The 20-car model is developed using the moving vehicle model for the regular trainset. Three-dimensional frame elements are used for the bridge model. Using the developed models, a dynamic behavior analysis program is coded. The analytical results are compared with the dynamic field test results and found to be valid to yield quite accurate dynamic responses. Based on the results of this study, the hybrid model, made up of the moving vehicle model for the heaviest power car and the moving force model for the other cars, is quite simple and effective without loosing the accuracy that much. Under the coincidence condition of two trains traveling with resonance velocity in the opposite directions, it is necessary to check not only the dynamic responses of the bridge with one-way traffic but those with two- way coincidence.

• Journal of Korea Game Society
• /
• v.15 no.4
• /
• pp.69-78
• /
• 2015

As the online game industry has been growing rapidly, more and more malicious activities to gain economic benefits have been reported as well. Game bot is one of the biggest problems in the online game industry. So we proposed a bot detection method based on the ERG theory of motivation for the first time. Most of the previous studies focused on behavior-based detection by monitoring patterns of the specific actions. In this paper, we applied the motivation theory to analyze user behaviors on a real game dataset. The result shows that normal users in the game followed the ERG theory of motivation in the same way as it works in real world. But in the case of game bots, the theory could not be applied because the game bot has specific reasons, unlike normal game users. We applied the ERG theory to users to distinguish game bot users from normal users. We detected the game bot with high accuracy of 99.78% by applying the theory.

• Journal of the Society of Naval Architects of Korea
• /
• v.35 no.1
• /
• pp.88-97
• /
• 1998

Large aluminium extrusions can be used in rolling stocks and high speed vessels to reduce weight and labor cost. As sandwich plates with corrugated core have enough strength in transverse and longitudinal direction, welding lines to connect members are reduced and transverse members to strength longitudinal members are not required. However, for proper design of aluminium extrusion plates, understanding of structural behaviors of the exclusions are necessary. In this paper, at first, detailed finite element analysis is carried out to understand structural behavior. And then, simple theoretical formulas for design purpose are proposed using the orthotropic plate theory. Shear stresses resulting from end twisting which is characteristics of deep aluminium extrusion plates can also be calculated by the simple theoretical formula. Comparison with the results by detailed finite element analysis shows good accuracy of the proposed simple formulas. The simple formulas can be useful in repetitive analysis in the initial design stage.