• Journal of Internet Computing and Services
• /
• v.21 no.6
• /
• pp.23-31
• /
• 2020

This study was carried out to generate various images of railroad surfaces with random defects as training data to be better at the detection of defects. Defects on the surface of railroads are caused by various factors such as friction between track binding devices and adjacent tracks and can cause accidents such as broken rails, so railroad maintenance for defects is necessary. Therefore, various researches on defect detection and inspection using image processing or machine learning on railway surface images have been conducted to automate railroad inspection and to reduce railroad maintenance costs. In general, the performance of the image processing analysis method and machine learning technology is affected by the quantity and quality of data. For this reason, some researches require specific devices or vehicles to acquire images of the track surface at regular intervals to obtain a database of various railway surface images. On the contrary, in this study, in order to reduce and improve the operating cost of image acquisition, we constructed the 'Defective Railroad Surface Regeneration Model' by applying the methods presented in the related studies of the Generative Adversarial Network (GAN). Thus, we aimed to detect defects on railroad surface even without a dedicated database. This constructed model is designed to learn to generate the railroad surface combining the different railroad surface textures and the original surface, considering the ground truth of the railroad defects. The generated images of the railroad surface were used as training data in defect detection network, which is based on Fully Convolutional Network (FCN). To validate its performance, we clustered and divided the railroad data into three subsets, one subset as original railroad texture images and the remaining two subsets as another railroad surface texture images. In the first experiment, we used only original texture images for training sets in the defect detection model. And in the second experiment, we trained the generated images that were generated by combining the original images with a few railroad textures of the other images. Each defect detection model was evaluated in terms of 'intersection of union(IoU)' and F1-score measures with ground truths. As a result, the scores increased by about 10~15% when the generated images were used, compared to the case that only the original images were used. This proves that it is possible to detect defects by using the existing data and a few different texture images, even for the railroad surface images in which dedicated training database is not constructed.

• Journal of The Korean Association For Science Education
• /
• v.39 no.6
• /
• pp.799-807
• /
• 2019

The study was conducted to identify if the first-year program of science-focused school improved students' scientific core competency and science learning motivation. The first-year program of the science-focused school consisted of basic education in scientific inquiry, investigation, advanced experiment, and basic education in small research. There were a total of 262 participants in the program, and 169 students took three survey tests. Through the analysis of a group-based trajectory modeling, students were differentiated based on similarity of score changes. This study showed that the first-year program of the science-focused school significantly improved students' scientific core competency and science learning motivation. A group-based trajectory modeling found that about 40~60% of students saw the effects of the program. The students who chose the humanity track showed effects, while some students who chose the science-focused track did not show effects. A group-based trajectory modeling showed the methodological effects of identifying the change process of individual students. This study identified the positive effects of science-focused school policy statistically and is a meaningful example for analyzing the effectiveness of science-focused school programs.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.2
• /
• pp.112-118
• /
• 2009

This paper proposes a rapid detection algorithm of a salient region for music video browsing system, which can be applied to mobile device and digital video recorder (DVR). The input music video is decomposed into the music and video tracks. For the music track, the music highlight including musical chorus is detected based on structure analysis using energy-based peak position detection. Using the emotional models generated by SVM-AdaBoost learning algorithm, the music signal of the music videos is classified into one of the predefined emotional classes of the music automatically. For the video track, the face scene including the singer or actor/actress is detected based on a boosted cascade of simple features. Finally, the salient region is generated based on the alignment of boundaries of the music highlight and the visual face scene. First, the users select their favorite music videos from various music videos in the mobile devices or DVR with the information of a music video's emotion and thereafter they can browse the salient region with a length of 30-seconds using the proposed algorithm quickly. A mean opinion score (MOS) test with a database of 200 music videos is conducted to compare the detected salient region with the predefined manual part. The MOS test results show that the detected salient region using the proposed method performed much better than the predefined manual part without audiovisual processing.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.3
• /
• pp.587-592
• /
• 2020

This paper analyzed the curriculum of computer and software-related majors educating the core IT-related skills needed for the 4th Industrial Revolution. The analysis was conducted on 158 majors classified as applied software, computer science and computer engineering according to the standard classification of university education units by the Standard Classification Committee of the Korean Council of University Education. The current status of introduction of curricular divided into the fields of Internet of Things(IoT) & mobile, cloud & big data, artificial intelligence(AI), and information security was analyzed among the contents of education in the relevant departments. According to the analysis, an average of 81.6% of the majors for each group of curricular organized related subjects into the curriculum. The Curriculum Response Index for the 4th industrial revolution(CRI4th) by major, calculated by weighting track operations by education sector, averaged 27.5 point out of 100 point. And the IoT & mobile sector had the highest score of 42.3 points.

• Journal of the Korean Data and Information Science Society
• /
• v.21 no.6
• /
• pp.1031-1039
• /
• 2010

The decathlon is an athletic event consisting of ten track and field events. Events are held over two consecutive days and the winners are determined by the combined performance in all. Performance is judged in meters, centimeters, minutes, and seconds. However, how to convert results into points is a difficult and controversial issue. We explored the distribution of decathlon results from the 1991 to 2009 using top 200 decathlons in the Olympic games and word championships. The conclusion is that the results from top level decathlon competition are normally distributed, and the current scoring system does not have the property that the performance with same difficulty should get same points. A new model for evaluating the decathlon score has been applied that display uniform characteristics over all events in order to meet the notion of allroundness. The proposed model is uniform over the events and support self-stabilization.

• PNF and Movement
• /
• v.12 no.3
• /
• pp.143-150
• /
• 2014

Purpose: The corticospinal tract (CST) is known to be an important pyramidal tract for walking and motor function. However, very little is known about the functional role of the CST in the recovery of motor function. In the current study, we investigated the relation between the CST and motor function in chronic hemiparetic stroke patients. Methods: Fifty-four patients and 20 normal subjects were recruited. The Functional Ambulation Category (FAC) was used in measurement of the walking ability. We classified patients into three groups according to the ability to walk independently: group A, patients who could not walk independently (FAC: 0-2); group B, patients who could walk independently (FAC: 3); and group C, patd walk functionally (stairs and uneven surfaces, FAC 4-5). The Motricity Index (MI) was used to measure the motor function of the affected upper and lower extremities (maximum score: 100). The fractional anisotropy (FA) value, apparent diffusion coefficient (ADC) value, and fiber volume of the CST were used for the diffusion tensor imaging (DTI) parameters. Results: In terms of the CST of the unaffected hemisphere, the FA value of group A was significantly lower than that of normal controls (p <0.05). The fiber volume of group C was significantly higher than that of normal controls (p <0.05). In contrast, the ADC values of all patient groups and the control group did not show any difference (p >0.05). In terms of lower MI and total MI, significant differences were observed between all patient groups (p <0.05). In addition, significant differences in terms of the upper MI scores were observed between groups A and C and between groups B and C (p <0.05); however, no significant difference was observed between groups A and B (p>0.05). Conclusion: The increased fiber volume of the CST in the unaffected hemisphere appears to be related to functional walking ability in chronic stroke patients. This result would be useful for elucidation of the neural recovery mechanism of walking and the investigation of new modalities for the recovery of walking following a stroke with CST injury.

• International Journal of Computer Science & Network Security
• /
• v.21 no.3
• /
• pp.134-140
• /
• 2021

The development of technology speeds up the process of obtaining information and its analysis to track the level of corruption in different countries and develop countermeasures. This study examines the role of information and analytical support of anti-corruption policy as a tool for government accountability and analysis, evaluation, combating corruption in Eastern Europe. The purpose of the article is to identify the components of the information-analytical system that help reduce the level of corruption. The research methodology is based on a qualitative content analysis of the functioning of information and analytical systems of Ukraine used by anti-corruption bodies. A quantitative analysis of the CPI score was conducted, according to Transparency International, to identify the effectiveness of anti-corruption policies in developing countries. The results show similar trends in countries developing on the effect of the use of information and analytical systems in the implementation of anti-corruption policies, strategies and measures. The strategy to combat corruption mainly involves increasing the independence and powers of anti-corruption bodies. Therefore, the development of information and analytical support is aimed at automating the processes of pre-trial investigations and criminal proceedings, information protection. As a tool for accountability, information and analytical systems may be ineffective due to the abuse of power by higher anti-corruption bodies due to political pressure from elite structures. Restrictions on political will are a major problem for the effectiveness of anti-corruption policies.

• Journal of The Korean Association For Science Education
• /
• v.30 no.4
• /
• pp.452-471
• /
• 2010

As the Korean College scholastic Ability Test (CSAT) has been implemented for 17 years since 1994, it is becoming more and more difficult to make new items that haven't been previously used to measure students' thinking ability. Therefore, it is necessary to keep conducting research on making new test items that can measure students' scholastic ability reliably. For this reason, multiple choice items on the Japanese university entrance exam, which is a Japanese National Center Test for University Admissions (NCTUA) equivalent of CSAT, were analyzed in order to draw implications for CSAT item development. In this study, we analyzed the Japanese NCTUA administered in January 2009 to investigate the structure of its science test. We also analyzed the NCTUA items by the domains of contents and behaviors, and tried to predict item difficulty from the perspective of Korean applicants. Major findings are as follows: Most NCTUA items measure understanding knowledge or low level thinking ability. Also the alloted time for each item is longer than CSAT. The number of test items, and the number of choice and alloted points for each item are diverse, unlike CSAT. The number of items using real-life materials are much more, but the items are not rigorous in sentence expression compared to CSAT. And the difference of difficulty level among science tests were larger with reference to CSAT. Also science score is required for most applicants regardless whether they are taking liberal arts or going onto the science track.

• Progress in Medical Physics
• /
• v.15 no.4
• /
• pp.215-219
• /
• 2004

Due to their excellence for the high-energy therapy range of photon beams, researchers show increasing interest in applying MOSFET dosimeters to low- and medium-energy applications. In this energy range, however, MOSFET dosimeter is complicated by the fact that the interaction probability of photons shows significant dependence on the atomic number, Z, due to photoelectric effect. The objective of this study is to develop a very detailed 3-dimensional Monte Carlo simulation model of a MOSFET dosimeter for radiological characterizations and calibrations. The sensitive volume of the High-Sensitivity MOSFET dosimeter is very thin (1 ${\mu}{\textrm}{m}$) and the standard MCNP tallies do not accurately determine absorbed dose to the sensitive volume. Therefore, we need to score the energy deposition directly from electrons. The developed model was then used to study various radiological characteristics of the MOSFET dosimeter. the energy dependence was quantified for the energy range 15 keV to 6 MeV; finding maximum dependence of 6.6 at about 40 keV. A commercial computer code, Sabrina, was used to read the particle track information from an MCNP simulation and count the tracks of simulated electrons. The MOSFET dosimeter estimated the calibration factor by 1.16 when the dosimeter was at 15 cm depth in tissue phantom for 662 keV incident photons. Our results showed that the MOSFET dosimeter estimated by 1.11 for 1.25 MeV photons for the same condition.

• Korean Journal of Applied Biomechanics
• /
• v.17 no.3
• /
• pp.197-207
• /
• 2007

It has been reported that Back-Tension played a significant role in archery (Lee & Bondit, 2005; Kim, 2007) but there are a few researches related Back-Tension in Korea recently. Therefore, the purpose of this study was to investigate archery back tension technique for the second ranked archer in the World and to find ways to improve performance. A subject(height: 185cm, mass: 82kg, years: 21yrs, careers: 12yrs) who is a number of national team and the second ranked archer in the World authorized by FITA (Federation Internationale de Trial Arc) was perticipated in this experiment. When shooting 60 shots($12{\times}5$), shooting motions were recorded with 7 infrared cameras and 2 ultrahigh-speed cameras. A QTM and an Auto Track were used to acquire raw data. The sampling rates of both cameras were 200 Hz. and 1000 Hz. respectively and data were filtered using a fourth order Butterworth low pass filtering with a cutoff-frequency of 30Hz. The parameters were calculated with Matlab6.5 and analyzed with SPSS11.0. After Pearson's correlations between 8 parameters were analyzed, 5 parameters from 13parameters that affected records were analyzed with multiple regression analysis (Enter order: x1, x2, x3, x4, x5). The results were as follows: 1. Comparing between parameters according to scores, the patterns of horizontal and vertical angular velocity(av.) of scapular relative angle was different between 8 score and 9 or 10 scores. 2. The correlations of parameters that affected records were a horizontal av.(x1, p=.032<.05) and a vertical av.(x3, p=.033<.05) of scapular from release to delivery in KB back-tension (anchoring-delivery). 3. The decision coefficients(R2) of above two parameters and three parameters selected by experts that may affect record, that is, an absolute trunk angle(x4) from in KKC back-tension (anchoring-release) and a horizontal relative scapular angle(x2) and an absolute trunk angle(x5) from release to delivery in KB back-tension were 7.7%(x1), 0.1%(x2), 8.5%(x3), 0.7%(x4) and 0.9%(x5) in sequence. 4. The multiple regression equation was a y= -1.16E-2 x1 + 0.109 x2 + 3.437E-2 x3 + 6.139E-2 x4 + 0.117 x5 + 3.420 In conclusion, a total contribution was low, that is, R2(17.9%) suggested that on the one hand, Lim's motion may not depend on a certain factor because his postural factors affected shooting motion are some stable on the other hand, unknown factors may exist(e.g. psychological, physiological factors etc.). Further study of EMG patterns of muscles and anatomic consideration related to shoulder girdle and scapular bones may help to identify mechanism of Back-Tension.