• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.1
• /
• pp.201-220
• /
• 2020

Trackers, especially long-term (LT) trackers, now have a more complex structure and more intensive computation for nowadays' endless pursuit of high accuracy and robustness. However, computing efficiency of LT trackers cannot meet the real-time requirement in various real application scenarios. Considering heterogeneous CPU-GPU platforms have been more popular than ever, it is a challenge to exploit the computing capacity of heterogeneous platform to improve the efficiency of LT trackers for real-time requirement. This paper focuses on TLD, which is the first LT tracking framework, and proposes an efficient parallel implementation based on OpenCL. In this paper, we firstly make an analysis of the TLD tracker and then optimize the computing intensive kernels, including Fern Feature Extraction, Fern Classification, NCC Calculation, Overlaps Calculation, Positive and Negative Samples Extraction. Experimental results demonstrate that our efficient parallel TLD tracker outperforms the original TLD, achieving the 3.92 speedup on CPU and GPU. Moreover, the parallel TLD tracker can run 52.9 frames per second and meet the real-time requirement.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.14 no.2
• /
• pp.175-183
• /
• 2014

In this paper, we present the design and implementation of real-time sound localization based on $0.13{\mu}m$ CMOS process. Time delay of arrival (TDOA) estimation was used to obtain the direction of the sound signal. The sound localization chip consists of four modules: data buffering, short-term energy calculation, cross correlation, and azimuth calculation. Our chip achieved real-time processing speed with full range ($360^{\circ}$) using three microphones. Additionally, we developed a dedicated sound localization circuit (DSLC) system for measuring the accuracy of the sound localization chip. The DSLC system revealed that our chip gave reasonably accurate results in an experiment that was carried out in a noisy and reverberant environment. In addition, the performance of our chip was compared with those of other chip designs.

• Journal of the Korean Society of Visualization
• /
• v.2 no.2
• /
• pp.8-15
• /
• 2004

With all the recent progresses in computer hardware and software technology, the animation of fluids in real-time is still among the most challenging issues of computer graphics. The fluid animation is carried out in two steps - the physical simulation of fluids immediately followed by the visual rendering. The physical simulation is usually accomplished by numerical methods utilizing the particle dynamics equations as well as the fluid mechanics based on the Navier-Stokes equations. Particle dynamics method is usually fast in calculation, but the resulting fluid motion is conditionally unrealistic. The methods using Navier-Stokes equation, on the contrary, yield lifelike fluid motion when properly conditioned, yet the complexity of calculation restrains this method from being used in real-time applications. This article presents a rapid fluid animation method by using the continuum-based fluid mechanics and the enhanced particle dynamics equations. For real-time rendering, pre-integrated volume rendering technique was employed. The proposed method can create realistic fluid effects that can interact with the viewer in action, to be used in computer games, performances, installation arts, virtual reality and many similar multimedia applications.

• Bulletin of the Society of Naval Architects of Korea
• /
• v.23 no.2
• /
• pp.1-13
• /
• 1986

The nonlinear hydrodynamic forces acting on a two-dimensional circular cylinder, oscillating with large amplitude in the free surface, are calculated by using the Semi-Lagrangian Time-Step-ping Method used by O.M. Faltinsen. In present calculation the position and the potential value of free surface are calculated using the exact kinematic and dynamic free surface boundary condition. At each time step an integral equation is solved to obtain the value of potential and normal velocity along the boundaries, consisting of both the body surface and the free surface. Some effort was devoted to the elimination of instability arising in the range of high frequency. Numerical simulations were performed up to the 3rd or 4th period which seems to be enough for the transient effect to die out. Each harmonic component and time-mean force are obtained by the Fourier transform of forces in time domain. The results are compared with others' experimental and theoretical results. Particularly, the calculation shows the tendency that the acceleration-phase 1st-harmonic component(added mass) increases as the motion amplitude increases and a reverse tendency in the velocity-phase 1st-harmonic component(damping coefficient). The Yamashita's experimental result also shows the same tendency. In general, the present result show relatively good agreement with the Yamashita's experimental result except for the time-mean force.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2022.11a
• /
• pp.51-52
• /
• 2022

In previous studies, fire simulation was used to estimate the fire spread path. According to previous studies, the fire spread path was estimated to be the main staircase, but consideration of interior materials and internal bulkheads was insufficient. In this study, the ignition time of the 3rd layer was analyzed using the prediction formula considering the interior materials and internal bulkheads. As a result of referring to the architectural drawings, it was found that the interior material of the 3rd floor was made of polystyrene. The internal ignition time of the third floor using FDTs was calculated to be 14,070 seconds (about 234 minutes). The internal ignition time of the 3rd floor using the Handbook on Design Calculation Methods of Fire Behavior was calculated to be 3,104 seconds (about 51 minutes). As a result of calculating the ignition time through the predictive formula, there is a large difference in the ignition time, so it is necessary to review the condition of the variable as a result of the calculation in the future.

• Fire Science and Engineering
• /
• v.30 no.3
• /
• pp.73-78
• /
• 2016

The Installation, Maintence, and Safety Control of Fire-fighting Systems Act of South Korea regulates that over 30-storey high-rise buildings including underground spaces should vitally perform the Performance-based Design to minimize property damage and personal injury as a fire risk assessment in advance. Therefore a PBD designer such as a fire safety professional engineer evaluate occupant's life safety by a scientific methodology. In order to evaluate the life safety, fire safety designers calculate the Required Safety Egress Time (RSET) which does not have the legal criteria regarding the standard method of calculation yet. So this way has been showing different results depending upon the designer's choice, knowledges and experiences. In this study, RSET calculation methods by six designers respectively were analysed from the thirteen reports of real performance based design projects conducted in Busan for a last five years. In particular, the Response Time calculation methods which have the most powerful effect for figuring the RSET are compared with the other designer's to deduce an error value.

• IE interfaces
• /
• v.6 no.2
• /
• pp.91-101
• /
• 1993

Developed in this paper is a standard time data for shirt sewing operations in the apparel-making industry. Since this data is made of formula developed for each unit operation and includes the formula for the machine time calculation, users could use it for obtaining the normal or standard time without any special efforts. In addition, this data handles the variations due to the design changes of a shirt through the parameters in the formula. compared with the MTM-1. The standard time data developed is shown to be easier and faster in determining the standard time with acceptable accuracy.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.46 no.4
• /
• pp.283-289
• /
• 2018

Launch Acceptability Region(LAR) means an area for successfully hitting the target. And LAR should be calculated in real time on aircraft so that LAR can be seen by pilot. LAR can be changed by the launch condition of the bomb, the impact condition of the target, and the atmospheric condition at the time of flight of the bomb. In this paper, we propose the calculation method of LAR using Artificial Neural Network(ANN). The learning data was generated by changing each condition from existing LAR model, and LAR model was derived through ANN learning. We confirmed the accuracy of the new LAR model by comparing the difference between the result data of existing LAR model and the new LAR model. And we confirmed the possibility of real time calculation of the LAR model on the aircraft by comparing the calculation time.

• Korean Journal of Construction Engineering and Management
• /
• v.12 no.6
• /
• pp.120-129
• /
• 2011

An installation of the construction lift has a few limitations by many constrains and these have influences on labor productivity, that can be changed by vertical-transportation management. In the super-tall building construction, a management of construction lift operation is one of the most important factor, but existing methodologies depend on skilled practitioners' experiences. And it is true that the expertise resulted by the experiences does not transfer to the next generation. This study is a part of lifting-management simulation development which aims at the optimal construction lift management. A proposed algorithm is focus on lifting time calculation considering an acceleration capability. This research evaluates the result accuracy using comparative analysis on simulation result and field measuring time.

• Proceedings of the Korean Society of Medical Physics Conference
• /
• 2002.09a
• /
• pp.116-118
• /
• 2002

It has been noted that Monte Carlo simulations are the most accurate method to calculate dose distributions in any material and geometry. Monte Carlo transport algorithms determine the absorbed dose by following the path of representative particles as they travel through the medium. Accurate Monte Carlo dose calculations rely on detailed modeling of the radiation source. We modeled the effects of beam modifiers such as collimators, blocks, wedges, etc. of our accelerator, Varian Clinac 600C/D to ensure accurate representation of the radiation source using the EGSnrc based BEAM code. These were used in the EGSnrc based DOSXYZ code for the simulation of particles transport through a voxel based Cartesian coordinate system. Because Monte Carlo methods use particle-by-particle methods to simulate a radiation transport, more particle histories yield the better representation of the actual dose. But the prohibitively long time required to get high resolution and accuracy calculations has prevented the use of Monte Carlo methods in the actual clinical spots. Our ultimate aim is to develop a Monte Carlo dose calculation system designed specifically for radiation therapy planning, which is distinguished from current dose calculation methods. The purpose of this study in the present phase was to get dose calculation results corresponding to measurements within practical time limit. We used parallel processing and some variance reduction techniques, therefore reduced the computational time, preserving a good agreement between calculations of depth dose distributions and measurements within 5% deviations.