• The Journal of the Korea Contents Association
• /
• v.16 no.12
• /
• pp.132-142
• /
• 2016

The study aimed to examine the state of visual acuity tests in preschoolers and determine the factors affecting the deterioration of their vision. A total of 172 pair, child aged 5-7 years and either one of their parents, were recruited from childcare centers or kindergarten between August 2015 and January 2016. Questionnaires were administered to the parents, and the children's visual acuities were measured. Using forward selection in logistics regression analysis, factors affecting vision deterioration were elucidated. Of 172 children, 128 (74.4%) had undergone a visual acuity test previously and 77 (44.8%) showed a deterioration in their vision. Children who had undergone their first visual acuity test after the age of 7 years (OR=7.425, CI=2.844- 19.385) and showed more abnormal eye symptoms, such as squinting or tilting the head to see or falling down frequently (OR=3.084, CI=1.202-7.914) and whose age was younger (OR=10.335, CI=3.252-32.848), were more likely to develop deterioration of vision. Children who had a posture such that they looked up at the computer monitor from below (OR=.075, CI=.022-.255), were less likely to show deterioration of vision. It can be inferred that early visual acuity tests is essential to detect deteriorated vision of preschoolers. Therefore, it is necessary to educate parents, personnels in day care center, and health clinics about the importance of early visual acuity tests and close observation of specific behaviors related to vision deterioration. Development of precautionary intervention program of vision deterioration in preschoolers and examination of its effects are needed.

• Geophysics and Geophysical Exploration
• /
• v.12 no.4
• /
• pp.316-327
• /
• 2009

A standard for national seismological observatory was proposed on 1999. Since then, Quanterra digitizer has been installed and is operating on almost all of seismic stations which belong to major seismic monitoring organizations. Quanterra digitizer produce and transmit real-time event packet and data packet. Characteristics of event packet and arrival time of each channel's data packet on data center were investigated. Packet selection criteria using signal to noise ratio (hereafter SNR) and signal period from real-time event packet based on 100 samples per second (hereafter sps) velocity data were developed. Estimation of epicenter using time information of the selected event packet were performed and tested. A series of experiment show that event packets were received approximately 3~4 second earlier than data packets and the number of event packet was only 0.3% compare to data packets. Just about 5% against all of event packets were selected as event packet were related P wave of real earthquake. Using the selected event packets we can estimate an epicenter with misfit less than 10 km within 20 sec for local earthquake over magnitude 2.5.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.17 no.2
• /
• pp.15-23
• /
• 2017

The person Re-identification is the most challenging part of computer vision due to the significant changes in human pose and background clutter with occlusions. The picture from non-overlapping cameras enhance the difficulty to distinguish some person from the other. To reach a better performance match, most methods use feature selection and distance metrics separately to get discriminative representations and proper distance to describe the similarity between person and kind of ignoring some significant features. This situation has encouraged us to consider a novel method to deal with this problem. In this paper, we proposed an enhanced recurrent neural network with three-tier hierarchical network for person re-identification. Specifically, the proposed recurrent neural network (RNN) model contain an iterative expectation maximum (EM) algorithm and three-tier Hierarchical network to jointly learn both the discriminative features and metrics distance. The iterative EM algorithm can fully use of the feature extraction ability of convolutional neural network (CNN) which is in series before the RNN. By unsupervised learning, the EM framework can change the labels of the patches and train larger datasets. Through the three-tier hierarchical network, the convolutional neural network, recurrent network and pooling layer can jointly be a feature extractor to better train the network. The experimental result shows that comparing with other researchers' approaches in this field, this method also can get a competitive accuracy. The influence of different component of this method will be analyzed and evaluated in the future research.

• Korean Journal of Health Education and Promotion
• /
• v.26 no.4
• /
• pp.35-47
• /
• 2009

Objectives: The purpose of this research was to investigate incidence and related factors of the MS in a university hospital Methods: This survey was conducted from January in 1997 to March in 2008. This study is a cohort study. The baseline subjects were 529 men and 609 women without metabolic syndrome(MS) who visited the health promotion center of Chungnam National University Hospital for general health check-up. The MS was diagnosed according to the definition by the NCEP ATP III. The abdominal obesity guidelines for waist circumference applied the Korean Society for the Study of Obesity Guideline Results: The survey exhibited that baseline risk factors that mainly affect odds ratio(OR) of the MS at follow-up were sex, ages, heavy drinking, uric acid and transpeptidase(GGT) in women. Incidence of the MS was 10.6% in total cases, 18.1% in men and 4.0% in women. The risk of the MS was significantly lower in women than men(OR=0.4, 95% confidence interval(CI: 0.2-0.9), significantly higher 50-59 Ages than < 40 ages(OR=3.4, 95% CI: 1.9-6.4), significantly high in women with heavy drinkers(OR=14.3, 95% CI: 1.2-177.5), significantly high in whole, men and women with high uric acid group(whole, OR=2.6, 95% CI: 1.6-4.2. men, OR=2.1, 95% CI: 1.3-3.5. women, OR=14.2, 95% CI: 3.0-68.3) and significantly high in women with high GGT(OR=1.8, 95% CI: 2.4-12.6). Conclusion: Risk factors of incident MS were baseline sex, ages, heavy drinking in women, uric acid and GGT in women. These can be utilized as an important index to determine risk factors of incident MS and will be basic data in part of management, education, countermeasure and selection test of the MS.

• Journal of the Korean Institute of Gas
• /
• v.18 no.4
• /
• pp.76-85
• /
• 2014

To establish the necessary safety technology in high-pressure toxic gas facilities, especially for the corrosion, which is the main causes of toxic gas accident, this study adopts and investigates the API-581 procedures developed by the American Petroleum Institute (API). And the applicability of the 8-step analytical procedures of consequence analysis in API-581 is discussed, and a method for consequence analysis in high-pressure toxic gas facilities is suggested. Based on the discussion and results, the analytical procedure is simplified as the 6 steps in total for the effective application to high-pressure toxic gas facilities: Step 1 (determination of representative material), Step 5 (determination of release type), Step 6 (determination of phase of fluid), and Step 8 (estimation of damage range) are not applied: Step 3 (estimation of total amount of release) is applied only for the inventory group concept; Step 4 (estimation of release rate) only for the gas release rate; and all of Step 2 (selection of release hole size) and Step 7 (evaluation of post-release response) are applied. In the proposed method, the generally applicable method of CCPS is adopted as alternative method for Steps 5 and 8.

• Korean Chemical Engineering Research
• /
• v.51 no.5
• /
• pp.561-567
• /
• 2013

Mixed hardwood chips were pre-pulping extracted with green liquor prior to kraft pulping in order to recover hemicelluloses for use as biofuels. This green liquor solution containing mainly sodium sulfide and sodium carbonate was applied at different alkali charges (expressed as $Na_2O$) of 0, 1, 3, and 5% on dry wood weight. The extractions were performed at $160^{\circ}C$ for residence times ranging from about 1-2 h to determine the effect of extraction severity on composition of the pre-pulping extract. The severity of hemicellulose extraction time and green liquor charge controls the concentration of acetic acid and monosaccharide sugars available for downstream processing, the accumulation of degradation products such as organic acids and furans in the extract. As the alkali charge was increased, the amount of acetate side chains on the hemicelluloses and the dissolved lignin in the extract increased but the carbohydrate and sugars in the extract decreased appreciably. Hot water extraction (0% alkali addition) released the greatest amount of carbohydrates, up to 29.80 g/L measured as component sugars, but resulted in the greatest decrease in pulp yield. Meanwhile, pre-pulping extraction with 3% green liquor increased the pulp yield while greatly reducing the component sugars to 7.08 g/L. Fundamental data obtained in this study will allow selection of optimum hemicellulose extraction conditions for integrating the extraction operation into the Kraft pulping process.

• Asian Journal of Turfgrass Science
• /
• v.24 no.1
• /
• pp.62-66
• /
• 2010

Synthetic turfgrass was developed for longer durability than natural turfgrass. As synthetic turfgrass use increases, disadvantage was exposed and composite turfgrass was designed to reinforce disadvantage of synthetic turfgrass. However, A few researches were conducted to evaluate composite turfgrass in South Korea. Therefore, this research was conducted to select a turfgrass species to maximize practical use of composite turfgrass. In 14 Oct. 2006 synthetic turfgrass was established in the research center in Hanul Sports Turf, Inc. located Hapcheon-Gun, Gyunggnam province. Kentucky bluegrass, Tall fescue, and a mixture of Kentucky bluegrass and Perennial ryegrass were used to combine with synthetic turfgrass. Wide and narrow types of synthetic turfgrass were used. As temperature increase, coverage of tall fescue and the mixture reduced but Kentucky bluegrass had the best result of turfgrass coverage although there were no differences on turfgrass quality among types of turfgrass.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.20 no.4
• /
• pp.419-425
• /
• 2014

The current work investigated the variation of numerical solutions according to the grid number, the distance of the first grid point off the ship surface, turbulence modeling and discretization. The subject vessel is KVLCC. A commercial code, Gridgen V15 and FLUENT were used the generation of the ship hull surface and spatial system and flow computation. The first part of examination, the effect of solutions were accessed depending on the grid number, turbulence modeling and discretization. The second part was focus on the suitable selection of the distance of the first grid point off the ship surface: $Y_P+$. When grid number and discretization were fixed the same value, the friction resistance showed differences within 1 % but the pressure resistance showed big differences 9 % depending on the turbulence modeling. When $Y_P+$ were set 30 and 50 for the same discretization, friction resistance showed almost same results within 1 % according to the turbulence modeling. However, when $Y_P+$ were fixed 100, friction resistance showed more differences of 3 % compared to $Y_P+$ of 30 and 50. Whereas pressure resistance showed big differences of 10 % regardless of turbulence modeling. When turbulence modeling and discretization were set the same value, friction, pressure and total resistance showed almost same result within 0.3 % depending on the grid number. Lastly, When turbulence modeling and discretization were fixed the same value, the friction resistance showed differences within 5~8 % but the pressure resistance showed small differences depending on the $Y_P+$.

• Journal of The Korean Association of Information Education
• /
• v.14 no.3
• /
• pp.291-300
• /
• 2010

Group-based learning is known to be an effective means to improve scholastic achievement in online learning. Therefore, there are some previous researches for the group-based learning. A lot of previous researches define factors for grouping from the characteristics of classes, teacher's decision and students' preferences and then generate a group based on the defined factors. However, many algorithms proposed by previous researches depend on a specific class and is not a general approach since there exist several differences in terms of the need of courses, learners, and teachers. Moreover it is hard to find a automatic system for group generation. This paper proposes a grouping system which automatically generate a learner group according to characteristics of various classes. the proposed system automatically generates a learner group by using basic information for a class or additional factors inputted from a user. The proposed system defines a set of rules for learner grouping which enables automatic selection of a learner grouping algorithm tailored to the characteristics of a given class. This rule based approach allows the proposed system to accommodate various learner grouping algorithms for a later use. Also we show the usability of our system by serviceability evaluation.

• Journal of Digital Convergence
• /
• v.11 no.10
• /
• pp.701-710
• /
• 2013

The purpose of this study is to lead the post digital generation of professional beautician training purpose is to find an efficient educational method. Beauty Education Facaility survey conducted in the center of reminders and educational statistics (