• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.2
• /
• pp.693-709
• /
• 2018

In this paper, we propose a compact Convolutional Neural Network (CNN)-based tracker in conjunction with a particle filter architecture, in which the CNN model operates as an accurate candidates estimator, while the particle filter predicts the target motion dynamics, lowering the overall number of calculations and refines the resulting target bounding box. Experiments were conducted on the Online Object Tracking Benchmark (OTB) [34] dataset and comparison analysis in respect to other state-of-art has been performed based on accuracy and precision, indicating that the proposed algorithm outperforms all state-of-the-art trackers included in the OTB dataset, specifically, TLD [16], MIL [1], SCM [36] and ASLA [15]. Also, a comprehensive speed performance analysis showed average frames per second (FPS) among the top-10 trackers from the OTB dataset [34].

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2001.04a
• /
• pp.475-480
• /
• 2001

The purpose of this study is to evaluate the validity of seismic performance evaluation using static analysis. For this purpose, Ordinary Moment Resisting Steel Frames(OMRSF) for different heights(3, 6 ,9, 12 story) and seismic zones(Zone 2A, 2B, 3, 4) were designed in compliance to AISC LRFD 1993 Seismic Provisions and NEHRP 1994 Guidelines. Nonlinear Static Procedure(NSP) and Nonlinear Dynamic Procedure(NDP) with a set of ground motion record were used to evaluate seismic demands in OMRSFs. Using the DRAIN-2DX program, this study compares peak displacement demands(Target Displacement) proposed by FEMA 273 with the peak roof displacement demands obtained from the inelastic time history analyses. Based on the results, the validity of procedure of seismic demand evaluation using Target Displacement is discussed.

• Journal of the Korean GEO-environmental Society
• /
• v.12 no.4
• /
• pp.53-63
• /
• 2011

Selection or generation of appropriate input ground motion is very important in performing a dynamic analysis. In Korea, it is a common practice to use recorded strong ground motions or artificial motions. The recorded motions show non-stationary characteristics, which is a distinct property of all earthquake motions, but have the problem of not matching the design response spectrum. The artificial motions match the design spectrum, but show stationary characteristics. This study generated ground motions that preserve the non-stationary characteristics of a real earthquake motion, but also matches the design spectrum. In the process, an impulse function-based algorithm that adjusts a given time series in time domain such that it matches the target response spectrum is used. Application of the algorithm showed that it can successfully adjust any recorded motions to match the target spectrum and also preserve the non-stationary characteristics. The modified motions are used to perform a series of nonlinear site response analyses. It is shown that the results using the adjusted motions result in more reliable estimates of ground vibration. It is thus recommended that the newly adjusted motions be used in practice instead of original recorded motions.

• Steel and Composite Structures
• /
• v.36 no.4
• /
• pp.399-415
• /
• 2020

Responses of semi-rigid frames having different degrees of semi-rigidity obtained by the nonlinear static analysis (NSA) are evaluated at specific target displacements by comparing them with those obtained by the nonlinear time-history analysis (NTHA) for scaled earthquakes. The peak ground accelerations (PGA) of the earthquakes are scaled such that the obtained peak top story displacements match with the target displacements. Three different types of earthquakes are considered, namely, far-field and near-field earthquakes with directivity and fling-step effects. In order to make the study a comprehensive one, three degrees of semi-rigidity (one fully rigid and the other two semi-rigid), and two frames having different heights are considered. An ensemble of five-time histories of ground motion is included in each type of earthquake. A large number of responses are considered in the study. They include the peak top-story displacement, maximum inter-story drift ratio, peak base shear, total number of plastic hinges, and square root of sum of the squares (SRSS) of the maximum plastic hinge rotations. Results of the study indicate that the nonlinear static analysis provides a fairly good estimate of the peak values of top-story displacements, inter-story drift ratio (for shorter frame), peak base shear and number of plastic hinges; however, the SRSS of maximum plastic hinge rotations in semi-rigid frames are considerably more in the nonlinear static analysis as compared to the nonlinear time history analysis.

• Progress in Medical Physics
• /
• v.28 no.4
• /
• pp.181-189
• /
• 2017

The conventional delivery quality assurance (DQA) process for RapidArc (Varian Medical Systems, Palo Alto, USA), has the limitation that it measures and analyzes the dose in a phantom material and cannot analyze the dosimetric changes under the motional organ condition. In this study, a DQA method was designed to overcome the limitations of the conventional DQA process for internal target volume (ITV) based RapidArc. The dynamic DQA measurement device was designed with a moving phantom that can simulate variable target motions. The dose distribution in the real volume of the target and organ-at-risk (OAR)s were reconstructed using 3DVH with the ArcCHECK (SunNuclear, Melbourne, USA) measurement data under the dynamic condition. A total of 10 ITV-based RapidArc plans for liver-cancer patients were analyzed with the designed dynamic DQA process. The average pass rate of gamma evaluation was $81.55{\pm}9.48%$ when the DQA dose was measured in the respiratory moving condition of the patient. Appropriate method was applied to correct the effect of moving phantom structures in the dose calculation, and DVH data of the real volume of target and OARs were created with the recalculated dose by the 3DVH program. We confirmed the valid dose coverage of a real target volume in the ITV-based RapidArc. The variable difference of the DVH of the OARs showed that dose variation can occur differently according to the location, shape, size and motion range of the target. The DQA process devised in this study can effectively evaluate the DVH of the real volume of the target and OARs in a respiratory moving condition in addition to the simple verification of the accuracy of the treatment machine. This can be helpful to predict the prognosis of treatment by the accurate dose analysis in the real target and OARs.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1202-1205
• /
• 2005

This paper describes a method for vision-based person identification that can detect, track, and recognize person from video using multiple cues: height and dressing colors. The method does not require constrained target's pose or fully frontal face image to identify the person. First, the system, which is connected to a pan-tilt-zoom camera, detects target using motion detection and human cardboard model. The system keeps tracking the moving target while it is trying to identify whether it is a human and identify who it is among the registered persons in the database. To segment the moving target from the background scene, we employ a version of background subtraction technique and some spatial filtering. Once the target is segmented, we then align the target with the generic human cardboard model to verify whether the detected target is a human. If the target is identified as a human, the card board model is also used to segment the body parts to obtain some salient features such as head, torso, and legs. The whole body silhouette is also analyzed to obtain the target's shape information such as height and slimness. We then use these multiple cues (at present, we uses shirt color, trousers color, and body height) to recognize the target using a supervised self-organization process. We preliminary tested the system on a set of 5 subjects with multiple clothes. The recognition rate is 100% if the person is wearing the clothes that were learned before. In case a person wears new dresses the system fail to identify. This means height is not enough to classify persons. We plan to extend the work by adding more cues such as skin color, and face recognition by utilizing the zoom capability of the camera to obtain high resolution view of face; then, evaluate the system with more subjects.

• Progress in Medical Physics
• /
• v.30 no.3
• /
• pp.65-73
• /
• 2019

Purpose: We evaluated the motion-induced dosimetric effects on the field-in-field (FIF) technique for whole-breast irradiation (WBI) using actual patient organ motion data obtained from cine electronic portal imaging device (cine EPID) images during treatment. Materials and Methods: Ten breast cancer patients who received WBI after breast-conserving surgery were selected. The static FIF (SFIF) plan involved the application of two parallel opposing tangential and boost FIFs. To obtain the amplitude of the internal organ motion during treatment, cine EPID images were acquired five times for each patient. The outside contour of the breast (OCB) and chest wall (CW) contour were tracked using in-house motion analysis software. Intrafractional organ motion was analyzed. The dynamic FIF (DFIF) reflecting intrafractional organ motion incorporated into the SFIF plan was calculated and compared with the SFIF in terms of the dose homogeneity index (DHI_90/10) for the target and V₂₀ for the ipsilateral lung. Results: The average motion amplitudes along the X and Y directions were 1.84±1.09 mm and 0.69±0.50 mm for OCB and 1.88±1.07 mm and 1.66±1.49 mm for CW, respectively. The maximum motion amplitudes along the X and Y directions were 5.53 and 2.08 mm for OCB and 5.22 and 6.79 mm for CW, respectively. Significant differences in DHI_90/10 values were observed between SFIF and DFIF (0.94 vs 0.95, P<0.05) in statistical analysis. The average V₂₀ for the lung in the DFIF was slightly higher than that of the SFIF in statistical analysis (19.21 vs 19.00, P<0.05). Conclusion: Our findings indicate that the FIF technique can form a safe and effective treatment method for WBI. Regular monitoring using cine EPID images can be effective in reducing motion-induced dosimetric errors.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2001.09a
• /
• pp.67-74
• /
• 2001

This paper presents the Safe Shutdown Earthquake(SSE) input motion for the seismic design of the Advanced Power Reactor 1400(APR1400). The Design Ground Response Spectra(DGRS) far the SSE is based on the design spectrum specified in regulatory Guide(RG) 1.60 of U.S. Nuclear Regulatory Commission(US NRC), anchored to a Peak Ground Acceleration(PGA) of 0.3g and enriched in the high frequency range. This SSE seismic input motion is to be applied to the seismic analysis as the free-field seismic motion at the ground surface of both the rock and generic soil sites fur APRI1400. The enrichment for APR1400 seismic input motion is performed considering the current US NRC regulations, the seismic hazard studies performed by the Lawrence Livermore National Laboratory (LINL) and Electric Power Research Institute(EPRI) for the Central and Eastern United States nuclear power plant sites, and the seismic input motions used in the design certifications of the three existing U.S. advanced standard plants. It is represented by a set of DGRS and the accompanying Target Power Spectral Density(PSD) Function in both the horizontal and vertical directions.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.19 no.3
• /
• pp.395-403
• /
• 2016

A cylinder-type underwater vehicle for military use that is running near the free surface at the final homing stage to hit a surface ship target is affected by wave force and moment. Since wave can affect an underwater vehicle running at the depth less than half of the modal wave length, it is important to confirm that the underwater vehicle can work well in such a situation. In this paper, wave force and moment per unit wave amplitude depending on wave frequency, wave direction, and vehicle's running depth were calculated by 3-Dimensional panel method, and the numerical results were modeled in external force terms of six degrees of freedom equations of motion. Motion simulation of the underwater vehicle running in various speed, depth, and sea state were performed.

• The Journal of the Acoustical Society of Korea
• /
• v.13 no.2
• /
• pp.30-37
• /
• 1994

Since the morden targets are of high speed and getting quiet in both active and passive mode, the necessities of developing advanced SONAR system capable of performing target motion analysis (TMA) and target classification are evident. In order to develop such a system, the scattering mechanism of complex bodies needs to be, some extent, fully understood and modeled. In this paper, MOving Spread Target(MOST) signal simulation model is presented and discussed. The model is based on the highlight distribution method, and simulates pulse elongation of spread target, doppler effect due to kinematics of the target as well as SONAR platform, and distribution target strength of each highlight point (HL) with directivity. The model can be used in developing and evaluating advanced SONAR system through system simulation, and can also be used in the development of target state estimation algorithm.