• Journal of Positioning, Navigation, and Timing
• /
• v.12 no.2
• /
• pp.91-100
• /
• 2023

The existing indoor localization method using Wi-Fi fingerprinting has a high collection cost and relatively low accuracy, thus requiring integrated correction of convergence with other technologies. This paper proposes a new method that significantly reduces collection costs compared to existing methods using Wi-Fi fingerprinting. Furthermore, it does not require labeling of data at collection and can estimate pedestrian travel paths even in large indoor spaces. The proposed pedestrian movement path estimation process is as follows. Data collection is accomplished by setting up a feature area near an indoor space intersection, moving through the set feature areas, and then collecting data without labels. The collected data are processed using Kernel Linear Discriminant Analysis (KLDA) and the valley point of the Euclidean distance value between two data is obtained within the feature space of the data. We build learning data by labeling data corresponding to valley points and some nearby data by feature area numbers, and labeling data between valley points and other valley points as path data between each corresponding feature area. Finally, for testing, data are collected randomly through indoor space, KLDA is applied as previous data to build test data, the K-Nearest Neighbor (K-NN) algorithm is applied, and the path of movement of test data is estimated by applying a correction algorithm to estimate only routes that can be reached from the most recently estimated location. The estimation results verified the accuracy by comparing the true paths in indoor space with those estimated by the proposed method and achieved approximately 90.8% and 81.4% accuracy in two experimental spaces, respectively.

• Journal of Information Technology Services
• /
• v.22 no.2
• /
• pp.1-21
• /
• 2023

Due to the convergence and complexity of the 4th Industrial Revolution, the boundaries between industries have become unclear and ambiguous. Consequently, there is a lack of research on how firms engaged in this industry are changing their location behavior. Recently, some attempts to classify the industrial groups of the 4th Industrial Revolution and their detail occupations have been made, and this study adopts the classification of Lee and Jung (2020) of the Korea Institute for Industrial Economics & Trade. In this study, the 18 detailed industries commonly included in multiple industrial groups are defined as 'core industries' and are classified into manufacturing and service industries to explore the spatial patterns of firms' location. Specifically, this study aims to examine how the location behavior of firms in core industries of the 4th Industrial Revolution has changed from 2010 to 2019 in the Seoul metropolitan area, using the 「National Business Survey」 data. We employed two methods based on spatial auto-correlation: (i) spatial kernel density estimation analysis and (ii) local Moran's I_i analysis. The results indicate that the core industry firms form more distinct and larger clusters in 2019 based on the clusters formed in 2010. Specifically, manufacturing industry firms tended to concentrate in the southern region of Gyeonggi and parts of Seoul, while serivce industry firms were more concentrated in Seoul. These core industries play a critical role in industries and are closely related to the ICT industries, which generate high-added value and increase productivity in the front and rear industries. This study reveals that the agglomeration of these industries in specific regions is intensifying and may exacerbate regional inequality.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.5
• /
• pp.1431-1445
• /
• 2022

We construct a deep neural network model named ECGResNet. This model can diagnosis diseases based on 12-lead ECG data of eight common cardiovascular diseases with a high accuracy. We chose the 16 Blocks of ResNet50 as the main body of the model and added the Squeeze-and-Excitation module to learn the data information between channels adaptively. We modified the first convolutional layer of ResNet50 which has a convolutional kernel of 7 to a superposition of convolutional kernels of 8 and 16 as our feature extraction method. This way allows the model to focus on the overall trend of the ECG signal while also noticing subtle changes. The model further improves the accuracy of cardiovascular and cerebrovascular disease classification by using a fully connected layer that integrates factors such as gender and age. The ECGResNet model adds Dropout layers to both the residual block and SE module of ResNet50, further avoiding the phenomenon of model overfitting. The model was eventually trained using a five-fold cross-validation and Flooding training method, with an accuracy of 95% on the test set and an F1-score of 0.841.We design a new deep neural network, innovate a multi-scale feature extraction method, and apply the SE module to extract features of ECG data.

• Journal of Internet of Things and Convergence
• /
• v.9 no.3
• /
• pp.21-26
• /
• 2023

The development and distribution approach of applications is transitioning from a monolithic architecture to microservices and containerization, a lightweight virtualization technology, is becoming a core IT technology. However, unlike traditional virtual machines based on hypervisors, container technology does not provide concrete security boundaries as it shares the same kernel. According to various preceding studies, there are many security vulnerabilities in most container images that are currently shared. Accordingly, attackers may attempt exploitation by using security vulnerabilities, which may seriously affect the system environment. Therefore, in this study, we propose an efficient automated deployment pipeline design to prevent the distribution of container images with security vulnerabilities, aiming to provide a secure container environment. Through this approach, we can ensure a safe container environment.

• The Korean Journal of Applied Statistics
• /
• v.37 no.1
• /
• pp.13-38
• /
• 2024

As economic uncertainties have increased recently due to COVID-19, there is a growing need to quickly grasp private consumption trends that directly reflect the economic situation of private economic entities. This study proposes a method of estimating private consumption in real-time by comprehensively utilizing big data as well as existing macroeconomic indicators. In particular, it is intended to improve the accuracy of private consumption estimation by comparing and analyzing various machine learning methods that are capable of fitting ultra-high-dimensional big data. As a result of the empirical analysis, it has been demonstrated that when the number of covariates including big data is large, variables can be selected in advance and used for model fit to improve private consumption prediction performance. In addition, as the inclusion of big data greatly improves the predictive performance of private consumption after COVID-19, the benefit of big data that reflects new information in a timely manner has been shown to increase when economic uncertainty is high.

• Annals of Occupational and Environmental Medicine
• /
• v.35
• /
• pp.23.1-23.14
• /
• 2023

Background: Exposure to heavy metals is a public health concern worldwide. Previous studies on the association between heavy metal exposure and neurobehavioral functions in children have focused on single exposures and clinical manifestations. However, the present study evaluated the effects of heavy metal complex exposure on subclinical neurobehavioral function using a Korean Computerized Neurobehavior Test (KCNT). Methods: Urinary mercury, lead, cadmium analyses as well as symbol digit substitution (SDS) and choice reaction time (CRT) tests of the KCNT were conducted in children aged between 10 and 12 years. Reaction time and urinary heavy metal levels were analyzed using partial correlation, linear regression, Bayesian kernel machine regression (BKMR), the weighted quantile sum (WQS) regression and quantile G-computation analysis. Results: Participants of 203 SDS tests and 198 CRT tests were analyzed, excluding poor cooperation and inappropriate urine sample. Partial correlation analysis revealed no association between neurobehavioral function and exposure to individual heavy metals. The result of multiple linear regression shows significant positive association between urinary lead, mercury, and CRT. BMKR, WQS regression and quantile G-computation analysis showed a statistically significant positive association between complex urinary heavy metal concentrations, especially lead and mercury, and reaction time. Conclusions: Assuming complex exposures, urinary heavy metal concentrations showed a statistically significant positive association with CRT. These results suggest that heavy metal complex exposure during childhood should be evaluated and managed strictly.

• Korean Journal of Radiology
• /
• v.21 no.7
• /
• pp.891-899
• /
• 2020

Objective: To assess the diagnostic performance of a deep learning-based algorithm for automated detection of acute and chronic rib fractures on whole-body trauma CT. Materials and Methods: We retrospectively identified all whole-body trauma CT scans referred from the emergency department of our hospital from January to December 2018 (n = 511). Scans were categorized as positive (n = 159) or negative (n = 352) for rib fractures according to the clinically approved written CT reports, which served as the index test. The bone kernel series (1.5-mm slice thickness) served as an input for a detection prototype algorithm trained to detect both acute and chronic rib fractures based on a deep convolutional neural network. It had previously been trained on an independent sample from eight other institutions (n = 11455). Results: All CTs except one were successfully processed (510/511). The algorithm achieved a sensitivity of 87.4% and specificity of 91.5% on a per-examination level [per CT scan: rib fracture(s): yes/no]. There were 0.16 false-positives per examination (= 81/510). On a per-finding level, there were 587 true-positive findings (sensitivity: 65.7%) and 307 false-negatives. Furthermore, 97 true rib fractures were detected that were not mentioned in the written CT reports. A major factor associated with correct detection was displacement. Conclusion: We found good performance of a deep learning-based prototype algorithm detecting rib fractures on trauma CT on a per-examination level at a low rate of false-positives per case. A potential area for clinical application is its use as a screening tool to avoid false-negative radiology reports.

• Journal of Practical Agriculture & Fisheries Research
• /
• v.20 no.1
• /
• pp.5-18
• /
• 2018

The field trial was performed to evaluate the rice growth and yield in direct seeding cultivation with iron-coated rice seeds. The required time for seed emergence was for 9~11days in the tested direct seeding methods. That was 1~2days earlier in direct seeding with pregerminated seeds than that of direct seeding with iron-coated seeds. The seedling establishment was highest in water seeding with iron-coated seeds but there was not significant difference in terms of statistical analysis. The rice plant height was taller in water seeding with broadcasting method than that of wet hill-seeding methods and in direct seeding with iron-coated seeds than that of direct seeding with pregerminated seeds. The tiller number in the rice plant was the highest in machine transplanting at 30days after direct seeding(June 17) and in water seeding with iron-coated seeds at 45days after seeding(DAS) and 60DAS. The tiller number of 75 and 90DAS in the tested rice cultivation methods being with 352~405/m² was not significantly different in terms of statistical analysis. The heading time was not different in rice direct seeding methods but 2 day earlier in direct seeding with iron-coated seeds than that of direct seeding with pregerminated seeds. The culm length was the highest in water seeding with iron-coated seeds and the panicle length was the longest in wet hill-seeding with pregerminated seeds. The panicle number per m² was highest in water seeding with iron-coated seeds but not significant difference among the tested rice cultivation methods. The water seeding with iron-coated seeds resulted in the highest spikelet number per m² and the heaviest grain weight of brown rice. Percentage of ripened kernel was the highest in wet hill-seeding with iron-coated seeds. But there were not significant among the tested rice cultivation methods. The milled rice yield in direct seeding methods was 3~21% higher than that in machine transplanting. Water seeding with iron-coated seeds recorded the highest milled rice yield being with 6.86t/ha.The occurrence of sheath blight was high according to machine transplanting>wet hill-seeding>water seeding. Weed occurrence was the highest in water seeding with pregerminated seeds. Weedy rice occurred not in machine transplanting but occured 0.6~0.7% in direct seeding methods with pregerminated seeds and 0.1% in direct seeding with iron-coated seeds.

• Journal of the Korean earth science society
• /
• v.26 no.7
• /
• pp.683-690
• /
• 2005

As remote sensing imagery with high spatial resolution (e.g. pixel resolution of 1m or less) is used widely in the specific application domains, the requirements of advanced methods for this imagery are increasing. Among many applicable methods, the texture image analysis, which was characterized by the spatial distribution of the gray levels in a neighborhood, can be regarded as one useful method. In the texture image, we compared and analyzed different results according to various directions, kernel sizes, and parameter types for the GLCM algorithm. Then, we studied spatial feature characteristics within each result image. In addition, a template matching program which can search spatial patterns using template images selected from original and texture images was also embodied and applied. Probabilities were examined on the basis of the results. These results would anticipate effective applications for detecting and analyzing specific shaped geological or other complex features using high spatial resolution imagery.

• Korean Security Journal
• /
• no.15
• /
• pp.199-219
• /
• 2008

This study analyzed the contents of the 225 papers included in Korea Security Science Association during the decade -from 1997 to 2007. This study was classified the study method qualitative. First, characteristic of researchers(distinction of sex, distinction of academic degree, regional distribution, one's position and regional distribution, participants per paper). Second, study trends classified by fields of study(where receiving research expenses support or not, change of study subject). Third, study trends classified by methods of study(study method by year, study method by study subject, statistical analysis by year) were subdivided. Analysis shows that there are some shortcomings on the research of Korea Security Science Association as compared with other fields. However, it shows advanced trends for example participation in different study field, evenly distributed regional study participation, variety trial of analysis method. Then again, the distinction of sex, one's position, too much emphasis on independence research, vulnerability about support of research expenses, emphasis on study fields and study trends wandering from industrial circles are getting deeper In study methods, generalized research form such as document study and phenomenon technical case study is limited so deduction of kernel result is not thoroughgoing enough as well as it shows the trend that limits to duplicate and generalized proposal technic.