KJC Korea
Journal Central

OpenGL compute shader is a shader stage that operate differently from other shader stage and it can be used for the calculating purpose of any data in parallel. This paper proposes a GPU-based parallel algorithm for computing sparse linear systems through conjugate gradient using an iterative method, which perform calculation on OpenGL compute shader. Basically, this sparse linear solver is used to solve large linear systems such as symmetric positive definite matrix. Four well-known matrix formats (Dense, COO, ELL and CSR) have been used for matrix storage. The performance comparison from our experimental tests using eight sparse matrices shows that GPU-based linear solving system much faster than CPU-based linear solving system with the best average computing time 0.64ms in GPU-based and 15.37ms in CPU-based.

In this paper, we demonstrate a performance boost by parallelizing a purely functional bisimulation algorithm on a multicore processor machine. The key idea of this parallelization is exploiting the referential transparency of purely functional programs to minimize refactoring of the original implementation without any parallel constructs. Both original and parallel implementations are written in Haskell, a purely functional programming language. The change from the original program to the parallel program is minuscule, maintaining almost original structure of the program. Through benchmark, we show that the proposed parallelization doubles the performance of the bisimulation test compared to the original non-parallel implementation. We also shaw that similar performance boost is also possible for a memoized version of the bisimulation implementation.

Cultural assets excavated in historical areas have their own characteristics based on the background of the times, and it can be seen that their patterns and characteristics change little by little according to the history and the flow of the spreading area. Cultural properties excavated in some areas represent the culture of the time and some maintain their intact appearance, but most of them are damaged/lost or divided into parts, and many experts are mobilized to research the composition and repair the damaged parts. The purpose of this research is to learn patterns and characteristics of the past through artificial intelligence neural networks for such restoration research, and to restore the lost parts of the excavated cultural assets based on Generative Adversarial Network(GAN)[1]. The research is a process in which the rest of the damaged/lost parts are restored based on some of the cultural assets excavated based on the GAN. To recover some parts of dammed of cultural asset, through training with the 2D image of a complete cultural asset. This research is focused on how much recovered not only damaged parts but also reproduce colors and materials. Finally, through adopted this trained neural network to real damaged cultural, confirmed area of recovered area and limitation.

In this paper, we propose a new stacking ensemble framework for deep learning models which reflects the distribution of label embeddings. Our ensemble framework consists of two phases: training the baseline deep learning classifier, and training the sub-classifiers based on the clustering results of label embeddings. Our framework aims to divide a multi-class classification problem into small sub-problems based on the clustering results. The clustering is conducted on the label embeddings obtained from the weight of the last layer of the baseline classifier. After clustering, sub-classifiers are constructed to classify the sub-classes in each cluster. From the experimental results, we found that the label embeddings well reflect the relationships between classification labels, and our ensemble framework can improve the classification performance on a CIFAR 100 dataset.

This study is to design and to develop the educational contents to enhance artificial intelligence literacy. First, we designed artificial intelligence education contents and constructed education programs. The contents are composed of a total of 15 lectures in 8 AI domains. The contents contain the elements of knowledge-skill-attitude, and 5 learning steps. The developed contents were organized in the form of online materials and included simulations and worksheets to directly manipulate and explore the concepts and algorithms of AI. In addition, we provided evaluation questions for each content. To examine the suitability of content, we conducted a validity test for experts. As a result of the content validity test, the overall average was .71 or higher, and the CVI value of the class suitability was .82, indicating a high validity. We are expected to use the contents developed in this study as an effective program to improve AI literacy in university liberal arts education.

This paper has proposed an algorithm that detecting for dense small vehicle in large image efficiently. It is consisted of two Ensemble Deep-Learning Network algorithms based on Coarse to Fine method. The system can detect vehicle exactly on selected sub image. In the Coarse step, it can make Voting Space using the result of various Deep-Learning Network individually. To select sub-region, it makes Voting Map by to combine each Voting Space. In the Fine step, the sub-region selected in the Coarse step is transferred to final Deep-Learning Network. The sub-region can be defined by using dynamic windows. In this paper, pre-defined mapping table has used to define dynamic windows for perspective road image. Identity judgment of vehicle moving on each sub-region is determined by closest center point of bottom of the detected vehicle's box information. And it is tracked by vehicle's box information on the continuous images. The proposed algorithm has evaluated for performance of detection and cost in real time using day and night images captured by CCTV on the road.

Recently, multi-modal deep learning techniques that combine heterogeneous data for deep learning analysis have been utilized a lot. In particular, studies on the synthesis of Text to Image that automatically generate images from text are being actively conducted. Deep learning for image synthesis requires a vast amount of data consisting of pairs of images and text describing the image. Therefore, various data augmentation techniques have been devised to generate a large amount of data from small data. A number of text augmentation techniques based on synonym replacement have been proposed so far. However, these techniques have a common limitation in that there is a possibility of generating a incorrect text from the content of an image when replacing the synonym for a noun word. In this study, we propose a text augmentation method to replace words using word hierarchy information for noun words. Additionally, we performed experiments using MSCOCO data in order to evaluate the performance of the proposed methodology.

Recentlly neural network approach for solving a singular perturbed integro-differential boundary value problem have been researched. Especially the model of the feed-forward neural network to be trained by the back propagation algorithm with various learning algorithms were theoretically substantiated, and neural network models such as deep learning, transfer learning, federated learning are very rapidly evolving. The purpose of this paper is to study the approaching method for developing a neural network model with high accuracy and speed for solving singular perturbed problem along with asymptotic methods. In this paper, we propose a method that the simulation for the difference between result value of singular perturbed problem and unperturbed problem by using neural network approach equation. Also, we showed the efficiency of the neural network approach. As a result, the contribution of this paper is to show the possibility of simple neural network approach for singular perturbed problem solution efficiently.

Today, the restoration of cultural properties is done by applying the latest IT technology from relying on existing data and experts. However, there are cases where new data are released and the original restoration is incorrect. Also, sometimes it takes too long to restore. And there is a possibility that the results will be different than expected. Therefore, we aim to quickly restore cultural properties using DeepLearning. Recently, so the algorithm DcGAN made in GANs algorithm, and image creation, restoring sectors are constantly evolving. We try to find the optimal GAN algorithm for the restoration of cultural properties among various GAN algorithms. Because the GAN algorithm is used in various fields. In the field of restoring cultural properties, it will show that it can be applied in practice by obtaining meaningful results. As a result of experimenting with the DCGAN and Style GAN algorithms among the GAN algorithms, it was confirmed that the DCGAN algorithm generates a top image with a low resolution.

In this paper, we propose a cultural heritage mobile augmented reality application that allows visitors to experience artifacts by augmenting prototypes, audio, video, and text information of 3D graphic artifacts of museum cultural assets. By applying augmented reality technology to a cultural product, products can be recognized on mobile phones and various historical information can be received through interaction of digital artifacts, and information can be easily and quickly checked through augmented reality mobile digital contents regardless of time and place. Through this study, we contribute to the development of digital cultural contents via mobile augmented reality and the expansion of augmented reality contents according to the types of cultural heritage for use, such as education, industry, and tourism promotion.

In this paper, a vehicle license plate detection scheme is proposed that uses the spatial attention areas to detect accurately the license plates in various real-road situations. First, the previous WPOD-NET was analyzed, and its detection accuracy is evaluated as lower due to the unnecessary noises in the wide detection candidate areas. To resolve this problem, a vehicle license plate detection model is proposed that uses the candidate area of the license plate as a spatial attention areas. And we compared its performance to that of the WPOD-NET, together with the case of using the optimal spatial attention areas using the ground truth data. The experimental results show that the proposed model has about 20% higher detection accuracy than the original WPOD-NET since the proposed scheme uses tight detection candidate areas.

In this paper, we compare two access control mechanisms for D2D(Device-to-Device) systems in 5G wireless networks and propose an effective access control for 5G D2D networks. Currently, there is no specified access control for 5G D2D networks but there can be two access control approaches for 5G D2D networks. One is the UE-to-Network Relay based access control and the other is the Remote UE(User Equipment) based access control. The former is a UE-to-Network Relay carries out the access control check for 5G D2D networks but the latter is a Remote UE performs the access control check for 5G D2D networks. Through simulation and evaluation, we finally propose the Remote UE based access control for D2D systems in 5G wireless networks. The proposed approach minimizes signalling overhead between the UE-to-Network Relay and the Remote UE and more efficiently performs the access control check, when the access control functionalities are different from the UE-to-Network Relay in 5G D2D networks.

This paper proposes a novel method of caching contents that can be encoded into multiple quality levels in device-to-device (D2D)-assisted caching networks. Different from the existing caching schemes, the author allows caching fractions of an individual file and considers the self cache hit event, which the user can find the desired content in its device. The author analyzes the tradeoff between the quality of cached contents and the cache hit rate, and proposes the device caching method maximizing the expected quality that the user can enjoy. Depending on the parameter of the relationship between the quality and the file size, the optimal caching method can be obtained by solving the convex optimization problem and the DC programming problem. If the file size increases faster than the quality, the cached fractions of the contents continuously increase as the popularity grows. Meanwhile, if the file size increases slower than the quality, some of the high-popularity files are entirely cached but others are not cached at all.

Companies' efforts to find ways to reduce logistics costs for products and raw materials currently being brought in to produce products in all manufacturing processes are one of the biggest challenges, and the recent global recession has made logistics management even more important. This development technology limits the development of IN/OUT Bound truck logistics and dock management optimization system in the factory, especially by using UHF 900Mhz RFID radio frequency technology.

In this study, an augmented reality-based make-up tool was built to analyze the user's face shape based on face-type reference model data and to provide virtual makeup by providing face-type makeup. To analyze the face shape, first recognize the face from the image captured by the camera, then extract the features of the face contour area and use them as analysis properties. Next, the feature points of the extracted face contour area are normalized to compare with the contour area characteristics of each face reference model data. Face shape is predicted and analyzed using the distance difference between the feature points of the normalized contour area and the feature points of the each face-type reference model data. In augmented reality-based virtual makeup, in the image input from the camera, the face is recognized in real time to extract the features of each area of the face. Through the face-type analysis process, you can check the results of virtual makeup by providing makeup that matches the analyzed face shape. Through the proposed system, We expect cosmetics consumers to check the makeup design that suits them and have a convenient and impact on their decision to purchase cosmetics. It will also help you create an attractive self-image by applying facial makeup to your virtual self.

VMP(Variable Message Protocol) is bit-based variable message processing protocol that enables the sharing situation information in real time as a tactical datalink protocol for Korean Army. System A is currently under development and will be operated as an army system when its development is completed. In system A, the VMP processing terminal is mounted and the VMP is utilized for exchanging tactical information. System A can acquire situation information from mounted situation information acquisition system and share situation information with other system A by the VMP. In this paper, we propose a method of sharing situation information with system A and speed improvement methods of sharing situation information using VMP. As speed improvement methods of sharing situation information, this paper studied 'Removing Process of Sending VMP Observation Report', 'Adopting One-time Situation Information Send Button When Sharing A System Situation Information By VMP', 'Combination of Many VMP Messages Using Repeat Function Of Application Header'. And we conducted the experiment, the result was that the transmission speed of equipment to which the research method was applied was improved by 76.8% compared to the existing equipment.

Naval Combat Management System(a.k.a. CMS) is the core combat power of ROK Navy. CMS which has been localized since 1993 has been developed in various categories. However, in the characteristic of defense industry, CMS software has been rarely developed technically and structurally while the environment of computing system has been developed dramatically. A need for a new paradigm of CMS software development was raised. This paper suggests Naval Shield Component Platform(NSCP) as a standard interface architecture of CMS based on SOLID of OOP which is an advanced programming paradigm and introduce its functionality and feature. We expect NSCP’s higher reusability, concurrency and maintainability in CMS software development. As a future work, we are going to apply NSCP to the next CMS software development project and evaluate quantitative, qualitative method.

Recently, researches for military purposes such as precision tracking and mission completion using UAVs have been actively conducted. In particular, if the posture information of the leading UAV is estimated and the mission UAV uses this information to follow in stealth and complete its mission, the speed of the posture information estimation of the guide UAV must be processed in real time. Until recently, research has been conducted to accurately estimate the posture information of the leading UAV using image processing and Kalman filters, but there has been a problem in processing speed due to the sequential processing of the processing process. Therefore, in this study we propose a way to improve processing speed by applying methods that the image processing area is limited to the ROI area including the object, not the entire area, and the continuous processing is distributed to OpenMP-based multi-threads and processed in parallel with thread synchronization to estimate attitude information. Based on the experimental results, it was confirmed that real-time processing is possible by improving the processing speed by more than 45% compared to the basic processing, and thus the possibility of completing the mission can be increased by improving the tracking and estimating speed of the mission UAV.

As a big data is being used in various industries, big data market is expanding from hardware to infrastructure software to service software. Especially it is expanding into a huge platform market that provides applications for holistic and intuitive visualizations such as big data meaning interpretation understandability, and analysis results. Demand for big data extraction and analysis using social media such as SNS is very active not only for companies but also for individuals. However despite such high demand for the collection and analysis of social media data for user trend analysis and marketing, there is a lack of research to address the difficulty of dynamic interlocking and the complexity of building and operating software platforms due to the heterogeneity of various social media service interfaces. In this paper, we propose a method for developing a framework to operate the process from collection to extraction and classification of social media data. The proposed framework solves the problem of heterogeneous social media data collection channels through adapter patterns, and improves the accuracy of social topic extraction and classification through semantic association-based extraction techniques and topic association-based classification techniques.

In this paper, the author proposes a method for detecting modification of transmitted messages in C/C++ based Discrete Event System Specification (DEVS) simulation. When a message generated by a model instance is delivered to other model instances, it may be modified by some of the recipients. Such modifications may corrupt simulation results, which may lead to wrong decision making. In the proposed method, every model instance stores a copy of every transmitted message. Before the deletion of the transmitted message, the instance compares them. Once a modification has been detected, the method interrupt the current simulation run. The procedure is automatically performed by a simulator instance. Thus, the method does not require programmers to follow secure coding or to add specific codes in their models. The performance of the method is compared with a DEVS simulator.

This study investigated the purchase status, wearing status, purchase satisfaction, selection criteria, and improvement marin leports wear for subjects who enjoy marine leisure sports in the southwestern region of Jeollanam-do, and examines the correlation between age and body type and the correlation between variables. SPSS Ver. 26.0 program was used for analysis. In the correlation between the motivation for participation and the selection criteria, when participating to increase physical strength, it was selected based on the fit. When participating for leisure or hobbies, they were selected based on design and color. The relationship between the selection criteria and purchase satisfaction was not satisfied in terms of price when they were selected based on activity or utilization of other uses. As for the selection criteria according to the body type, clothing was selected based on price and fashion for large triangles and squares, and elasticity for inverted triangles. As for the preferred color by age, only those in their 40s preferred blue and other age groups preferred achromatic color. In future studies, it is thought that a study on the preference of each marine leisure sports item should be conducted.

This study examines the relationship between internal marketing, market orientation, patient satisfaction, and hospital image, and especially focuses on the effect of CEO's role on internal marketing at a local national university hospital. A survey was conducted using the convenient sampling technique and 222 questionnaires excluding unreliable replies were used in the final analysis for the hypothesis testing. SPSS 21.0 was used for the basic analysis of the collected data, and confirmatory factor analysis was performed for reliability and validity using AMOS 21.0. Path analysis was performed for the hypothesis testing. The results of this study are as follows: First, the role of CEO positively affects internal marketing. Second, internal marketing has a positive effect on market orientation, and leadership is the most influential factor of internal marketing. Third, market orientation has a positive effect on patient satisfaction and hospital image, which are non-financial organizational performance. Therefore, internal marketing plays a major role in improving market orientation, patient satisfaction, and hospital image, and it is identified that the activation of internal marketing depends on the support of CEO in hospitals.

This study aims to provide the relevant basis upon which prediction of dyslipidemia should be made based on body information. Using the National Health Insurance big data (3,312,971 people) canonical correlation analysis was performed between body information and lipid-profile. Body information included age, height, weight and waist circumference, while the lipid-profile included total cholesterol, triglycerides, HDL cholesterol and LDL cholesterol. As a result, when the waist circumference and the weight are large, triglycerides increase and HDL cholesterol level decreases. In terms of age, weight, waist circumference, and HDL cholesterol, the canonical variates (the degree of influence) were significantly different according to sex. In particular, the canonical variate was dramatically changed around the forties and fifties in women in terms of weight, waist circumference, and HDL cholesterol. The canonical correlation results of the health care big data presented in this study will help construct a predictive model that can evaluate an individual's health status based on body information that can be easily measured in a non-invasive manner.

In this paper, I developed a scenario-based admission management virtual reality (SAM VR) content for practical training for nursing students and verified the effectiveness. The SAM VR contents used in the study was developed by the researcher using Gear VR and smartphone according to the standard practical procedure suggested by the Korea Acreditation Board of Nursing Education and Evaluation. In the 30 experimental groups who received practical training using SAM VR contents, learning flow, learning confidence, and learning satisfaction increased statistically significantly after the practical training (p<.001). In the control group, who received practical training in the traditional way, learning confidence increased after the practical training (p<.005), but there was no change in learning flow and learning satisfaction (p>.005). It was verified that the SAM VR contents are effective practical education contents for nursing students' learning flow, learning confidence and learning satisfaction.

Recently, trust in online transactions and corporate trust are most important at the corporate level as social overhead capital in commercial transactions using digital devices such as online, mobile, and SNS platforms. Therefore, this study used data from the Korea Consumer Agency's consumer policy indicators to identify the impact of digital consumption trust and corporate responsibility trust on the satisfaction of information and communication products and services. According to the analysis, trust in digital consumption conditions and responsibility of companies have a positive impact on satisfaction of information and communication devices and satisfaction of information and communication services. In addition, it was found that trust in corporate responsibility has a greater impact on satisfaction of information and communication devices and satisfaction of information and communication services than trust in digital consumption conditions. Theoretical and practical implications for these findings and suggestions for future research were presented.

The purpose of this study was to provide some policy implications by analyzing impacts of disability occurrence on economic, social cultural exclusion and testing empirically moderation effects of social environment in their relationships. For this study, 112 disabled persons were sampled and surveyed on economic, social cultural exclusion by disability. Analytical results are the followings. First, new disability caused economic exclusion. Second, it also caused social, cultural exclusion. Third, social support has a strong moderation effects between disability and economic, social cultural exclusion. Social support played the important role for reducing the negative impacts of disability occurrence on social exclusion. This study provided some policy implications about raising social support for the disabled basing on this analytical results. First, programs for improving cognition for the disabled need to be developed and implemented from the early childhood. Second, social campaign for the disabled are promoted positively by non-governmental sector. Third, public policy for the disabled should be strengthened from material support to even emotional support.

In this paper, we suggest a method of analysis for daily food delivery & consumption trends through big data of the post-Covid-19 era. Through analysis of big data and the database system, four analyzed factors, excluding weather, was proved to have significant correlation with delivery sales for ‘Baedarui Minjok’ of a catering delivery application. The research found that KBS, MBC and SBS Media showed remarkable results in food delivery & consumption sales soaring up to about 60 percent increase on the day after the Covid-19 related new article was issued. In addition, it proved that mobile media and web surfing were the main factors in increasing sales of food delivery & consumption applications, suggesting that viral marketing and emotional analysis by crawling data from SNS used by Millennials might be an important factor in sales growth. It can contribute the companies in the economic recession era to survive by providing the method for analyzing the big data and increasing their sales.

Traditional Chinese medicine has treated diseases and improved health in nature-based experience. Advanced nations began to be interested in naturopathic therapy in the late 19th century and it led China to research aromatherapy. This study searched previous researches related with aromatherapy and generally analyzed aroma oil, applied body parts, methods of use, and period of use. For research contents, scientific and society journals from 2000 to 2019 related with aromatherapy were searched in CNKI(www.cnki.com) and WANFANG DATE(www.wanfang.com). Finally, 30 papers were selected through 5-step qualitative evaluation and expert review and analyzed. Frequency and percentage(%) were calculated by means of the Excel 2013 Program and represented by a chart. The results of analyzing aromatherapy trends are as follows. All 30 papers were researched in the medical society. The most common symptom was irritation and anxiety that appeared in 13 papers. Lavender oil and bergamot oil were commonly used aroma oil. Commonly applied part and method were nose and nasal inhalation. For aroma oil associated with symptoms, lavender oil was the best in irritative, anxious, and negative emotion, depression, labor pain, sleep disorder, migraine, tension, and vomiting, pain, and fatigue after operation. Lemon, ginger, and peppermint oil was good for nausea. Based on the findings, this study derived applied body parts, methods of use, and period of use in aromatherapy. However, most aromatherapy was used for patients in the nursing and medical fields in the simple form of inhalation and local massage. This study will suggest a standard ground that aromatherapy is good for pain, colic pain, and tension in a short period but needs a long period for the efficacy of psychological and neurological symptoms.

In this study, we explored consumers' the privacy attitudes and Internet activities on the expectations of the fourth industrial revolution. Furthermore, we examined the moderating effect of Internet activities between the privacy attitudes and the expectations of the fourth industrial revolution. Research data are 2018 Korea media panel survey provided by Korea Information Society Development Institute. As a result, concerns about privacy were negative on the expectations of the fourth industrial revolution. Consumer’s internet activities had a positive effect on the expectations of the fourth industrial revolution. Concerns about privacy have a negative effect on the expectations of the fourth industrial revolution, but active internet activities have mitigated it or turned it into a positive. This study will contribute as basic data for more active responses in the economic structure facing the 4th industrial revolution.

SNS has recently reached the level of providing financial services to customers through a mobile payment system that goes beyond the existing payment system using Fintech, which is a fusion of financial industry and information technology. These mobile payment systems are increasing in scale as time goes by, and their functions are reaching the same level as general financial services. This study is an empirical study to examine what is the most important factor in Internet banking by targeting users who use WeChat Pay among Chinese Internet bank users with the highest Fintech Adoption rate. SNS has recently reached the level of providing financial services to customers through a mobile payment system that goes beyond the existing payment system using Fintech, which is a fusion of financial industry and information technology. As a results, 2 factors positive influence on Acceptance intention and Customer satisfaction. These mobile payment systems are increasing in scale as time goes by, and their functions are reaching the same level as general financial services.

In this paper, we propose a novel recommendation model based on Doc2vec using search keywords and product details. Until now, a lot of prior studies on recommender systems have proposed collaborative filtering (CF) as the main algorithm for recommendation, which uses only structured input data such as customers’ purchase history or ratings. However, the use of unstructured data like online customer review in CF may lead to better recommendation. Under this background, we propose to use search keyword data and product detail information, which are seldom used in previous studies, for product recommendation. The proposed model makes recommendation by using CF which simultaneously considers ratings, search keywords and detailed information of the products purchased by customers. To extract quantitative patterns from these unstructured data, Doc2vec is applied. As a result of the experiment, the proposed model was found to outperform the conventional recommendation model. In addition, it was confirmed that search keywords and product details had a significant effect on recommendation. This study has academic significance in that it tries to apply the customers' online behavior information to the recommendation system and that it mitigates the cold start problem, which is one of the critical limitations of CF.

The purpose of this research is to explore what factors affect consumers' purchasing decisions when purchasing artificial intelligence artworks. The research pointed out that in the real shopping model, customer perceived value includes three dimensions: product perceived value, service perceived value and social perceived value. On this basis, an artificial intelligence work purchase decision-making influence model was constructed, and an online survey was attempted to collect data. Through analysis of the reliability, effectiveness and structural equations of SPSS24.0 and AMOS24.0, and scientific verification and analysis, we found that product cognitive value and service cognitive value have a positive impact on consumers’ purchase intentions, but social cognition Value has no positive effect on consumers' purchasing intentions.

In this paper, we propose the development and application of a software education program for underachievers. The software education program for underachieving students was developed in consideration of the characteristics of learner’s suffering from underachievement and the educational effects of software education, and is meaningful in that it proposes a plan to improve the learning gap in distance learning. Learners can acquire digital literacy and learning skills by solving structured tasks in the form of courseware, intelligent tutoring, debugging, and artificial intelligence learning models in educational programs. Based on the effects of software education, such as enhancing logical thinking ability and problem solving ability, this program provides opportunities to solve fusion tasks to underachievers. Based on this, it is expected that it can have a positive effect on the overall academic work.