• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.38 no.6
• /
• pp.70-81
• /
• 2001

This paper presents a 3D face modeling method using a CCD camera and a projector (LCD projector or Slide projector). The camera faces the human face and the projector casts white stripe patterns on the human face. The 3D shape of the face is extracted from spatial and temporal locations of the white stripe patterns on a series of image frames. The proposed method employs region segmentation and multi-beam techniques for efficient 3D modeling of hair region and faster 3D scanning respectively. In the proposed method, each image is segmented into face, hair, and shadow regions, which are independently processed to obtain the optimum results for each region. The multi-beam method， which uses a number of equally spaced stripe patterns, reduces the total number of image frames and consequently the overall data acquisition time. Light beam calibration is adopted for efficient light plane measurement, which is not influenced by the direction (vertical or horizontal) of the stripe patterns. Experimental results show that the proposed method provides a favorable 3D face modeling results, including the hair region.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.17 no.1
• /
• pp.45-55
• /
• 2021

Recently, competition in the interior remodeling business is fierce. However, the interior remodeling business is basically B2B, and in this business, sales are basically dictated by the capabilities of salespeople. Therefore, in this study, we intend to develop a model for the core competency of salespeople in the interior remodeling business. To this end, based on the research on the existing competency and competency modeling, the competency model of the salesperson was derived using Dubois' overlay method. A total of 12 core competencies could be defined through the first and second modeling. The subject of this study was focused on LG Hausys, the most representative interior remodeling company in Korea. Based on the core competencies developed in this way, overall sales competencies can be raised through the development of a training course to enhance the sales competencies of salespeople, and a more efficient and objective HR.

• ETRI Journal
• /
• v.32 no.5
• /
• pp.810-818
• /
• 2010

This paper presents a new framework for integrating untranscribed spoken content into the acoustic training of an automatic speech recognition system. Untranscribed spoken content plays a very important role for under-resourced languages because the production of manually transcribed speech databases still represents a very expensive and time-consuming task. We proposed two new methods as part of the training framework. The first method focuses on combining initial acoustic models using a data-driven metric. The second method proposes an improved acoustic training procedure based on unsupervised transcriptions, in which word endings were modified by broad phonetic classes. The training framework was applied to baseline acoustic models using untranscribed spoken content from parliamentary debates. We include three types of acoustic models in the evaluation: baseline, reference content, and framework content models. The best overall result of 18.02% word error rate was achieved with the third type. This result demonstrates statistically significant improvement over the baseline and reference acoustic models.

• Journal of Communications and Networks
• /
• v.16 no.4
• /
• pp.436-446
• /
• 2014

The high intensity of research and modeling in fields of mathematics, physics, biology and chemistry requires new computing resources. For the big computational complexity of such tasks computing time is large and costly. The most efficient way to increase efficiency is to adopt parallel principles. Purpose of this paper is to present the issue of parallel computing with emphasis on the analysis of parallel systems, the impact of communication delays on their efficiency and on overall execution time. Paper focuses is on finite algorithms for solving systems of linear equations, namely the matrix manipulation (Gauss elimination method, GEM). Algorithms are designed for architectures with shared memory (open multiprocessing, openMP), distributed-memory (message passing interface, MPI) and for their combination (MPI + openMP). The properties of the algorithms were analytically determined and they were experimentally verified. The conclusions are drawn for theory and practice.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.5
• /
• pp.472-482
• /
• 2008

A modulation transfer function(MTF) tree is established to estimate the overall MTF of an observation satellite and to analyze the image performance. Basic MTF models relevant to each MTF tree component are represented as mathematical relationship between optics-structural dynamics, thermal deformation, attitude and dynamic characteristics of a satellite and the effects due to the space environment. The Basic MTF models consist of diffraction limited MTF with central obscuration, aberration, defocus, line-of-sight(LOS) jitter, linear motion, detector integration, and so forth. Performance estimation is demonstrated for a virtual earth-observation satellite in order to validate the constructed modeling method. The proposed models enable the system engineers to calculate the overall system MTF and to determine the crucial design parameters that affect the image performance in the conceptual design phase of an observation satellite.

• International Journal of Internet, Broadcasting and Communication
• /
• v.11 no.3
• /
• pp.49-58
• /
• 2019

Efficient evidential reasoning is an important issue in the development of advanced knowledge based systems. Efficiency is closely related to the design of problems solving methods adopted in the system. The explicit modeling of problem-solving structures is suggested for efficient and effective reasoning. It is pointed out that the problem-solving method framework is often too coarse-grained and too abstract to specify the detailed design and implementation of a reasoning system. Therefore, as a key step in developing a new reasoning scheme based on properties of the problem, the problem-solving method framework is expanded by introducing finer grained problem-solving primitives and defining an overall control structure in terms of these primitives. Once the individual components of the control structure are defined in terms of problem solving primitives, the overall control algorithm for the reasoning system can be represented in terms of a finite state diagram.

• Land and Housing Review
• /
• v.11 no.2
• /
• pp.95-104
• /
• 2020

This study aims to explore a purely mechanistic pavement analysis approach where viscoelasticity and fracture of asphalt mixtures are considered to accurately predict deformation and damage behavior of flexible pavements. To do so, the viscoelastic and fracture properties of designated pavement materials are obtained through experiments and a fully mechanistic damage analysis is carried out using a finite element method (FEM). While modeling crack development can be done in various ways, this study uses the cohesive zone approach, which is a well-known fracture mechanics approach to efficiently model crack initiation and propagation. Different pavement configurations and traffic loads are considered based on three main functional classes of roads suggested by FHWA i.e., arterial, collector and local. For each road type, three different material combinations for asphalt concrete (AC) and base layers are considered to study damage behavior of pavement. A concept of the approach is presented and a case study where three different material combinations for AC and base layers are considered is exemplified to investigate progressive damage behavior of pavements when mixture properties and layer configurations were altered. Overall, it can be concluded that mechanistic pavement modeling attempted in this study could differentiate the performance of pavement sections due to varying design inputs. The promising results, although limited yet to be considered a fully practical method, infer that a few mixture tests can be integrated with the finite element modeling of the mixture tests and subsequent structural modeling of pavements to better design mixtures and pavements in a purely mechanistic manner.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 1999.10a
• /
• pp.365-373
• /
• 1999

Recently, numerous studies have demonstrated that artificial intelligence such as neural networks can be an alternative methodology for classification problems to which traditional statistical methods have long been applied. In building neural network model, the selection of independent and dependent variables should be approached with great care and should be treated as a model construction process. Irrespective of the efficiency of a learning procedure in terms of convergence, generalization and stability, the ultimate performance of the estimator will depend on the relevance of the selected input variables and the quality of the data used. Approaches developed in statistical methods such as correlation analysis and stepwise selection method are often very useful. These methods, however, may not be the optimal ones for the development of neural network models. In this paper, we propose a genetic algorithms approach to find an optimal or near optimal input variables for neural network modeling. The proposed approach is demonstrated by applications to bankruptcy prediction modeling. Our experimental results show that this approach increases overall classification accuracy rate significantly.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.21 no.8
• /
• pp.1332-1338
• /
• 1997

In the case of the transfer function for the secondary path is dependent on time, the on-line method which can model it is continuously must be applied to the active noise control technique. And the adaptive random noise technique among the on-line methods is effective in the narrow-band control. In this method, the signal to noise ratio between random noise for modeling and primary noise is low. Therefore, the estimations of transfer function will be prone to inaccuracies and the convergence time will be too long. Such imperfections will have an influence upon the performance of an active noise controller. In this study, t enhance the signal to noise ratio, the on-line method that is combined the conventional adaptive random noise technique and the adaptive line enhancer, is proposed. By using proposed on-line method, a rigorous system identification and control of primary noise have been implemented.

• Journal of Soil and Groundwater Environment
• /
• v.28 no.6
• /
• pp.9-15
• /
• 2023

In order to properly evaluate the spatio-temporal variations of groundwater flow, the data obtained in field experiments should be corroborated into numerical simulations. Particle tracking method is a simple simulation tool often employed in groundwater simulation to predict groundwater flow paths or solute transport paths. Particle tracking simulations visually show overall the particle flow path along the entire aquifer, but no previous simulation studies has yet described the parameter values at grid nodes around the particle path. Therefore, in this study, a new technical approach was proposed that enables acquisition of parameters associated with particle transport in grid nodes distributed in the center of the particle path in groundwater. Since the particle tracking path is commonly referred to as streamline, the algorithm and codes developed in this works designated streamline analysis method. The streamline analysis method can be applied in two-dimensional and three-dimensional finite element or finite difference grid networks, and can be utilized not only in the groundwater field but also in all fields that perform numerical modeling.