KJC Korea
Journal Central

In this paper, we propose a training algorithm of support vector machine (SVM) with a sensitive variable. Although machine learning models enable automatic decision making in the real world applications, regulations prohibit sensitive information from being used to protect privacy. In particular, the privacy protection of the legally protected attributes such as race, gender, and disability is compulsory. We present an efficient least square SVM (LSSVM) training algorithm using a fully homomorphic encryption (FHE) to protect a partial sensitive attribute. Our framework posits that data owner has both non-sensitive attributes and a sensitive attribute while machine learning service provider (MLSP) can get non-sensitive attributes and an encrypted sensitive attribute. As a result, data owner can obtain the encrypted model parameters without exposing their sensitive information to MLSP. In the inference phase, both non-sensitive attributes and a sensitive attribute are encrypted, and all computations should be conducted on encrypted domain. Through the experiments on real data, we identify that our proposed method enables to implement privacy-preserving sensitive LSSVM with FHE that has comparable performance with the original LSSVM algorithm. In addition, we demonstrate that the efficient sensitive LSSVM with FHE significantly improves the computational cost with a small degradation of performance.

An abnormal object refers to a person, an object, or a mechanical device that performs abnormal and unusual behavior and needs observation or supervision. In order to detect this through artificial intelligence algorithm without continuous human intervention, a method of observing the specificity of temporal features using optical flow technique is widely used. In this study, an abnormal situation is identified by learning an algorithm that translates an input image frame to an optical flow image using a Generative Adversarial Network (GAN). In particular, we propose a technique that improves the pre-processing process to exclude unnecessary outliers and the post-processing process to increase the accuracy of identification in the test dataset after learning to improve the performance of the model's abnormal behavior identification. UCSD Pedestrian and UMN Unusual Crowd Activity were used as training datasets to detect abnormal behavior. For the proposed method, the frame-level AUC 0.9450 and EER 0.1317 were shown in the UCSD Ped2 dataset, which shows performance improvement compared to the models in the previous studies.

This study proposes a deep learning-based flat foot classification methodology using transfer learning. We used a transfer learning with VGG16 pre-trained model and a data augmentation technique to generate a model with high predictive accuracy from a total of 176 image data consisting of 88 flat feet and 88 normal feet. To evaluate the performance of the proposed model, we performed an experiment comparing the prediction accuracy of the basic CNN-based model and the prediction model derived through the proposed methodology. In the case of the basic CNN model, the training accuracy was 77.27%, the validation accuracy was 61.36%, and the test accuracy was 59.09%. Meanwhile, in the case of our proposed model, the training accuracy was 94.32%, the validation accuracy was 86.36%, and the test accuracy was 84.09%, indicating that the accuracy of our model was significantly higher than that of the basic CNN model.

In this paper, we propose a deep learning-based retinal vessel segmentation model for handling multi-scale information of fundus images. we integrate the selective kernel convolution into U-Net-based convolutional neural network. The proposed model extracts and segment features information with various shapes and sizes of retinal blood vessels, which is important information for diagnosing eye-related diseases from fundus images. The proposed model consists of standard convolutions and selective kernel convolutions. While the standard convolutional layer extracts information through the same size kernel size, The selective kernel convolution extracts information from branches with various kernel sizes and combines them by adaptively adjusting them through split-attention. To evaluate the performance of the proposed model, we used the DRIVE and CHASE DB1 datasets and the proposed model showed F1 score of 82.91% and 81.71% on both datasets respectively, confirming that the proposed model is effective in segmenting retinal blood vessels.

In this paper, we propose a method for the motion recognition method required in the industrial field in mixed reality. In industrial sites, movements (grasping, lifting, and carrying) are required throughout the upper full-body, from trunk movements to arm movements. In this paper, we use a method composed of sensors and wearable devices that are not vision-based such as Kinect without using heavy motion capture equipment. We used two IMU sensors for the trunk and shoulder movement, and used Myo arm band for the arm movements. Real-time data coming from a total of 4 are fused to enable motion recognition for the entire upper body area. As an experimental method, a sensor was attached to the actual clothes, and objects were manipulated through synchronization. As a result, the method using the synchronization method has no errors in large and small operations. Finally, through the performance evaluation, the average result was 50 frames for single-handed operation on the HoloLens and 60 frames for both-handed operation.

In this paper, we propose a quadtree-based optimization technique that enables fast Super-resolution(SR) computation by efficiently classifying and dividing physics-based simulation data required to calculate SR. The proposed method reduces the time required for quadtree computation by downscaling the smoke simulation data used as input data. By binarizing the density of the smoke in this process, a quadtree is constructed while mitigating the problem of numerical loss of density in the downscaling process. The data used for training is the COCO 2017 Dataset, and the artificial neural network uses a VGG19-based network. In order to prevent data loss when passing through the convolutional layer, similar to the residual method, the output value of the previous layer is added and learned. In the case of smoke, the proposed method achieved a speed improvement of about 15 to 18 times compared to the previous approach.

In this paper, we propose a human-content interface that allows users to quickly and efficiently search data through friction-based scrolling with ROI(Regions of interests). Our approach, conceived from the behavior of finding information or content of interest to users, efficiently calculates ROI for a given content. Based on the kernel developed by conceiving from GMM(Gaussian mixture model), information is searched by moving the screen smoothly and quickly to the location of the information of interest to the user. In this paper, linear interpolation is applied to make one softer inertia, and this is applied to scrolls. As a result, unlike the existing approach in which information is searched according to the user's input, our method can more easily and intuitively find information or content that the user is interested in through friction-based scrolling. For this reason, the user can save search time.

Location-based services (LBSs) are expected to process a large number of spatial queries, such as shortest path and k-nearest neighbor queries that arrive simultaneously at peak periods. Deploying more LBS servers to process these simultaneous spatial queries is a potential solution. However, this significantly increases service operating costs. Recently, batch processing solutions have been proposed to process a set of queries using shareable computation. In this study, we investigate the problem of batch processing moving k-nearest neighbor (MkNN) queries in dynamic spatial networks, where the travel time of each road segment changes frequently based on the traffic conditions. LBS servers based on one-query-at-a-time processing often fail to process simultaneous MkNN queries because of the significant number of redundant computations. We aim to improve the efficiency algorithmically by processing MkNN queries in batches and reusing sharable computations. Extensive evaluation using real-world roadmaps shows the superiority of our solution compared with state-of-the-art methods.

In this paper, we introduce InfoDID, a robust user information management service based on DID that manages user information reliably. Since blockchain technology provides an environment in which data can be handled transparently on a decentralized basis, various services using blockchain are currently being developed As the importance of user's personal information has recently emerged, the DID technology is receiving attention. The technology allows a user to control his or her information through decentralized identifiers, and various information management services are being tried based on the technology. Using blockchain-based DID technology, InfoDID reliably controls personal information requested frequently, helping users to provide their information to other services more conveniently. In addition, to support service continuity, InfoDID uses BR2K technique that provides robust execution of blockchain application services, so that even partial service failures can be systematically recovered. To facilitate this operation, we present a replication status monitoring tool that can continuously check the replication states of blockchain application services running in association with the BR2K technique such as InfoDID.

Some sections of the highway are closed due to disasters and accidents. In this situation, it analyzes what kind of situation occurs due to functional failure in the highway network. The domestic highway network can be expressed as a graph. Blocking some sections of the highway can turn into a national disaster. In this paper, we analyze the robustness of the domestic highway network. The robustness of expressways refers to the degree to which the traffic conditions of the domestic expressway network deteriorate due to the blockage of some sections. The greater the robustness, the smaller the effect of some blocking appears. This study is used to evaluate the congestion level of one section of the transportation network, and a value obtained by dividing the section traffic volume (V) by the section traffic volume (C) is used. This study analyzes the robustness of highways by using the actual traffic volume data of the departure and arrival points of domestic highways, and analyzes the changes in traffic volume due to partial blockage through experimental calculations. Although this analysis cannot reflect the exact reality of domestic highways, it is judged to be sufficient for the purpose of confirming the basic robustness of the overall network.

Despite the efforts of financial authorities in conducting the direct management and supervision of collection agents and bond-collecting guideline, the illegal and unfair collection of debts still exist. To effectively prevent such illegal and unfair debt collection activities, we need a method for strengthening the monitoring of illegal collection activities even with little manpower using technologies such as unstructured data machine learning. In this study, we propose a classification model for illegal debt collection that combine machine learning such as Support Vector Machine (SVM) with a rule-based technique that obtains the collection transcript of loan companies and converts them into text data to identify illegal activities. Moreover, the study also compares how accurate identification was made in accordance with the machine learning algorithm. The study shows that a case of using the combination of the rule-based illegal rules and machine learning for classification has higher accuracy than the classification model of the previous study that applied only machine learning. This study is the first attempt to classify illegalities by combining rule-based illegal detection rules with machine learning. If further research will be conducted to improve the model's completeness, it will greatly contribute in preventing consumer damage from illegal debt collection activities.

In the era of big data, interest in data is exploding. In particular, the development of the Internet and social media has led to the creation of new data, enabling the realization of the era of big data and artificial intelligence and opening a new chapter in convergence technology. Also, in the past, there are many demands for analysis of data that could not be handled by programs. In this paper, an analysis model was designed and verified for classification of unstructured data, which is often required in the era of big data. Data crawled DBPia's thesis summary, main words, and sub-keyword, and created a database using KoNLP’s data dictionary, and tokenized words through morpheme analysis. In addition, nouns were extracted using KAIST's 9 part-of-speech classification system, TF-IDF values were generated, and an analysis dataset was created by combining training data and Y values. Finally, The adequacy of classification was measured by applying three analysis algorithms(random forest, SVM, decision tree) to the generated analysis dataset. The classification model technique proposed in this paper can be usefully used in various fields such as civil complaint classification analysis and text-related analysis in addition to thesis classification.

In this paper, we propose an efficient dynamic workload balancing strategy which improves the performance of high-performance computing system. The key idea of this dynamic workload balancing strategy is to minimize execution time of each job and to maximize the system throughput by effectively using system resource such as CPU, memory. Also, this strategy dynamically allocates job by considering demanded memory size of executing job and workload status of each node. If an overload node occurs due to allocated job, the proposed scheme migrates job, executing in overload nodes, to another free nodes and reduces the waiting time and execution time of job by balancing workload of each node. Through simulation, we show that the proposed dynamic workload balancing strategy based on CPU, memory improves the performance of high-performance computing system compared to previous strategies.

In this study, I compared the pattern design method of classic shirts focused on the men's clothing consturction textbook. It was intended to select a shirt pattern drawing method suitable for men in their 20s by conducing a appearance evaluation with a 3D simulation program. Pattern drawing method, sizes, appearance evaluation, garment pressure. were evaluated, and SPSS 26.0 program was used for analysis. According to the Pattern drawing method, there were differences in application sizes, and there were many parts where designated sizes were applied. The shirt pattern is mostly the same for the front chest and back chest, front waist and back waist. However, if there is a front and back difference of the chest and the waist circumference, the fit was considered better. B pattern was evaluated as the best in appearance evaluation, color distribution, and mesh condition through 3D simulation, and B pattern was analyzed as the most suitable method for men in their 20s. Since this study applied the average sizes of the 7th Korean Human Body Dimension Survey report in 20s, it is thought that caution should be paid to apply them to all 20s. In the future, it is also thought that research on the actual fit and patterns of various body types and materials in their 20s should be carried out.

The purpose of this study was to examine the most suitable exercise for the elderly women by comparing the changes in body composition and physical function fitness after modern dance, aquarobic and combined exercise programs for the elderly women. The subjects of this study were 47 elderly women in J-do, chosen as MDG(n=13), AEG(n=11), CEG(n=11) and CG(n=12) for participated for 60 minutes/day and three times/wk for 12 weeks. The results of grip strength, chair stand, arm curl, 2 minute step, 244cm up and go was significantly than in the control group. Back scratch was decreased significantly after exercise only in the modern dance group and combined exercise group but there was no change in chair sit and reach. These results suggest that various exercises were found to play a positive role in maintaining and improving fitness for the elderly women. In addition, additional exercise programs are needed to improve flexibility and mobility of joints for the elderly. It is thought that it is important to voluntarily participate in a movement that suits your interests to make it sustainable.

The purpose of this study is to investigate how the application of a Thorax belt affects the stability of the spine in subjects with a wide infra-sternal angle. A total of 15 subject with wide infra-sternal angle participated in the experiment. Active Double leg lowering and active one-leg raising were performed with or without a thorax belt. Two spinal stability tests (active double lowering and active one-leg raising) performed with or without thorax belt application showed significant differences between each condition. Based on the results of present study, the application of a thorax belt is considered to be an effective therapeutic tool that can stabilize the spine to subjects with abnormally increased chest cage and spinal or trunk instability.

This study wanted to analyze the types of recognition of messages expressed in clothing for women in their 20s who wear a lot of clothing and fashion products with text. It was intended to provide basic data necessary for the production of typography clothing and fashion products by considering the subjective evaluation of how women in their 20s type the characters expressed in fashion and the characteristics of each type. This study was conducted with the Q method, and the QUANL pc program was used for analysis. Type I thought that letters were a design element and fashion, and the characters expressed in clothes were recognized as images. Type 2 thought it was important that the characters expressed in the clothing were recognized as messages, and that the characters had social messages and period reflections. Type 3 preferred that letters be combined with casual clothes and valued the formability of the characters. Type 4 preferred characters to represent brands and liked to be placed in large positions. In the future, it is thought that additional research by various age groups and genders and detailed research should be conducted to identify differences in font, color, and sentence length.

In this paper, it is a system that informs people who take medication periodically to facilitate the convenience of the elderly and the disabled. It is a system that measures the full weight of pills that need to be taken for a week using a weight sensor, and then determines whether or not the pills are taken by measuring the weight of the reduced pills again when the user takes them. For people with disabilities who are unable to move, it includes the function of automatically transporting medicine to the user-set location at the time of use using a line tracer based autonomous vehicle. It is also configured to inform users who have not taken the pill through an alarm that includes visual and auditory functions at a specific time to inform them of this. This work attempts to help users take their medication without forgetting by segmenting the task performance process of such a system through simulations.

The purpose of this study is to examine types and characteristics of self-regulation which explain how individual diabetic patients determine their behavior in terms of diabetes management. Based on the results, we proposed the management method of the diabetic patients. As a method of analyzing the subjectivity of each item, a Q method study was used. There were 38 subjects who were treated with oral drugs or insulin injections, but not currently hospitalized. And 40 statements in Q samples were classified in the form of a normal distribution using a 9-point scale. Research results regarding self-regulation in diabetic patients, type 1 is evidence-based compliance, type 2 is crisis responding denial reaction, type 3 is information-based orientation to relationship, type 4 is independent practice, and type 5 is willingness deficit impulse reaction. Therefore, for effective management of diabetic patients, it is necessary to understand the characteristics of each type of self-regulation and develop a program that reflects motivation for diabetes management, improvement of confidence, and countermeasures for negative emotions related to disease.

The purpose of the current study was to investigate the causal relationship between school life adjustment and life satisfaction among adolescents through longitudinal panel data. The current study analyzed the 1st through 7th wave data in the 4th grade panel of elementary school from the Korea Children and Youth Panel Survey (KCYPS). The research model was tested using auto-regressive cross-lagged model. The major results were as follows. First, adolescents’ school life adjustment had a positive auto-regressive effect. Second, adolescents’ life satisfaction had a positive auto-regressive effect. Third, adolescents’ school life adjustment was a causal predictor of life satisfaction, but not vice versa. Finally, implications in terms of enhancing adolescents’ school life adjustment and life satisfaction were also discussed.

The point of this research is verifing the longitudinal changes in youth multicultural acceptance and to verify predictive variables of multicultural acceptance. Among the Korean Children and Youth Panal data collected by the Korea Institute for Children and Youth Policy, 1,972 data from the 3rd, 5th, and 6th data of the first-year middle school panal were used, and the data were analyzed by applying the latent growth model, and the following analysis results were obtained. First, adolescents’ multicultural acceptability showed an increasing pattern, and the extent of the increase was large at the point of transition from the 3rd year(3rd year of middle school) to the 5th year. Second, predicting the change of multicultural acceptability were found to be significant in the initial values of peer communication, community consciousness and in the rate of change, life-satisfaction, peer communication, community consciousness. This study proposed some plans to improve multicultural acceptance in adolescence basing on these results.

The main purpose of this study is to check whether the family relationship of adolescents has an impact on smartphone addiction and to verify of direct effects and moderating effects in according to self- efficacy. The survey was conducted to the adolescents who were second grade middle school students with 318 subjects in G metropolitan city. This study results found the followings: Family relationship and selfe-efficacy revealed direct influence on smartphone addiction. Moreover, the self-efficacy check the moderating effect of do family relationship on smartphone addiction. The results of this study are significant in that they demonstrated the influence of family relationship and self-effect on smartphone addiction and imply that self-efficacy can decrease or prevent smartphone addiction from family relationship problem.

With the advent of the age of"untact" modern people are pursuing a healthy body and mind. In order to achieve well-being, LOHAS and Wellness,people prefer to use natural affinity alternative therapies, Aromatherapy. This study focuses on women in their 20s~50s in Zhejiang Province, with the aim of investigating their knowledge and use of essential oils.The questionnaire was divided into four parts: 3 questions for general question, 11 questions for knowledge, 13 questions for use and 9 questions for satisfaction. In addition, the study was conducted using the WeChat and the Wenjuanxing Program from July 5 to August 30, 2019. Finally, a total of 617 questionnaires were analyzed. In this study, SPSS WIN 21.0 program is used for frequency analysis. The level of knowledge and satisfaction is verified by Cronbach‘s α. And the following analysis results were obtained by frequency analysis, descriptive statistics, Chi-squared test(), one-way ANOVA on the understanding level and usege of essential oils according to age. The results were as follows. The most common characteristics of subjects were the 20s, university students, essential oil recognition was high in having experience. There is no great difference in knowledge or satisfaction depending on age. knowledge and satisfaction was moderate. The results of experience in the use of essential oils were higher among all age groups, those who in their 30s did not think that the use of essential oils would be effective. However, people in their 20s and 40s and older have unclear answers, indicating that results showed a difference. The results of the survey on usage showed that there were significant differences in period of use, place of purchase, method of purchase, purpose of use, place of use, number of use, frequency of use, body parts of use. According to the study, awareness and knowledge of essential oils vary according to age, and those in their 20s use essential oils for facial skin, and those in their 30s and older use essential oils for stress relief and body management. This study provides basic information on marketing related to diversified essential oil products according to age.

In this paper, we suggest the Deep Neural Network Model System for predicting results of the match of ‘League of Legends (LOL).’ The model utilized approximately 26,000 matches of the LOL game and Keras of Tensorflow. It performed an accuracy of 93.75% without overfitting disadvantage in predicting the ‘2020 League of Legends Worlds Championship’ utilizing the real data in the middle of the game. It employed functions of Sigmoid, Relu and Logcosh, for better performance. The experiments found that the four variables largely affected the accuracy of predicting the match --- ‘Dragon Gap’, ‘Level Gap’, ‘Blue Rift Heralds’, and ‘Tower Kills Gap,’ and ordinary users can also use the model to help develop game strategies by focusing on four elements. Furthermore, the model can be applied to predicting the match of E-sports professional leagues around the world and to the useful training indicators for professional teams, contributing to vitalization of E-sports.

This study investigated the determinants of debt ratio in public institutions. For this purpose, we analyzed the impact of inefficient management as internal factors on debt ratio. In this paper, inefficient management included total costs, payment, and employee benefit. The results of this study are as follows. First, we find that there is a significant positive relation between total costs and debt ratio. This result means that the higher total costs, the higher debt ratio. Second, we find that there is not a significant relation between payment and debt ratio. And we also find that there is not a significant relation between employee benefit and debt ratio. These results are empirical results that can be answers about some concerns that inefficient management of public institutions worsen debt ratio.

Over a decade ago, Krasnova et al. identified the factors that influence Facebook users’ self-disclosure. These factors include perceived risks, relationship building, relationship maintenance, self-presentation, and enjoyment. Meanwhile, during the past 10 years, there have been significant changes in terms of function, media, and competition. SNSs have been functionally enhanced, used in mobile environment, and had many competitors. Based on these facts, it is believed that the influence of the factors on self-disclosure is different from those of Krasnova et al. The purpose of this study is to verify through a replication study whether the factors adopted in the study of Krasnova et al. are still important in explaining self-exposure. The study empirically find the result significantly different from those of Krasnova et al. Based on the result, the study provides meaningful implications and suggestions for future research.

Due to the fourth industrial revolution and the Corona19, the availability of digital devices such as kiosks is a matter directly related to survival for older adults with digital divide. This paper analyzes the problems of existing information service education for older adults and proposes an IT education strategy tailored to aging and life cycle of older adults. The proposed in this study is an educational application-based IT training method that supports repetitive learning regardless of time and place, developing kiosk function as a functional game-type educational application and experimenting with post-training effectiveness. The research method proposed UI usability evaluation frame for older adults, and developed educational applications based on proposed evaluation frame, and conducted kiosk education. As a result of the experiment, the mission success rate after using the IT application was 80.6%, which is a 55.1%P improvement compared to the pre-use(25.5%). This study confirmed that the proposed education for older adults is a way to overcome the limitations of existing IT education in the current situation.