Abstract : Scoliosis is one of most common spine diseases. The measurement of Cobb angle is gold standard of the diagnosis of scoliosis. However, the reliability and accuracy of the cobb angle measurement has room for improvement. In this study, a model was trained for thoracic and lumbar vertebra segmentation based on Convolution Neural Networks. The data consisted of 252 data and 10-fold cross-validation was performed. Also, the measurement of cobb angle was performed using the deep learning-based approach and it was compared with that of manual measurement. As a result, the average sensitivity of the dataset was 0.927, the specificity was 0.999, the accuracy was 0.998, and the Dice’s similarity coefficient was 0.927. There is no statistical difference in cobb angle measured by the deep learning-based approach and the manual measurement. Also, the intraclass correlation coefficient between them was 0.973. Therefore, it is expected that vertebra segmentation using deep learning will be useful for the measurement of cobb angle.
Keyword : Artificial Intelligence, Deep learning, Scoliosis, Cobb angle, Segmentation
http://dx.doi.org/10.29056/jncist.2020.03.01
A Study of Technology and Application Plan on Context Detection for Customized Disaster
Abstract : Recent climate change and urban complication have dramatically increased the potential for natural andsocial disasters. As a result, as the interest in disaster response is focused, the goal of the disaster responseis to quickly recognize the disaster situation and apply a response manual for each disaster situation tominimize human and material damage. However, since the experience of the disaster management personnelis forced to accumulate while experiencing a disaster, there is a limit to recognizing the situation in theearly stage and performing an appropriate disaster response procedure. In this paper, we propose a deeplearning-based image recognition method for edge-based disaster situation awareness. According to thismethod, it is possible for disaster management personnel to quickly respond to a disaster by minimizingdisaster damage by real-time processing and analyzing on-site images to detect the occurrence of thedisaster and predict the situation of the risk and uncertainty.
Keyword : Disaster situationn transmission, edge computing, disaster customized, context detection
http://dx.doi.org/10.29056/jncist.2020.03.02
Abstract : This study aims at the realization of the right to be forgottenthe perspective of policy and industry. The goal is to eliminate the obstacles and problems of the realization of rights, and to suggest the direction of system improvement and service model development. I saw the realization of rights as the concept of the right to erasure and approachedtwo perspectives: information deletion and access exclusion. First, we analyzed the problems and obstacles focused on personal information processing. Second, we designed service considerations, goals, functional requirements, and use case diagrams for realizing rights. This study is a meaningful subject that has not existed before, and is expected to be the basis for related studies. The service will be completed through continuous research in the future.
Keyword : Right to be Forgotten, Right to Erasure, Privacy, Personal Information Protection(PIP), Right to Information Self-determination, Unified Modeling Language
http://dx.doi.org/10.29056/jncist.2020.03.03
A Study on Characteristics of Female Executives Who Run Small and Medium-size Enterprises in Korea
Abstract : Most prior studies are in analyzing the effects of characteristics of management policies about specific industries, regions, and corporate governance, and suggesting ways to improve the issues about the management activities of female managers among Korean companies. It is difficult to understand comprehensively the critical characteristics of women's management companies and present their future directions. This study identifies Korea women have some difficulties to operate start-up business because of national consciousness to sexual aspects of labor, fund raising, start-up item and business location In order to manifest the difficult conditions of women entrepreneurs and to promote their businesses, And this study suggests four governmental and political tactical approaches to enhance sustainability of women-owned company.
Keyword : Women Entrepreneurship, Entrepreneurial Fund Raising, Entrepreneurial Idea and item, Entrepreneurial Network, Entrepreneurial Sustainability
http://dx.doi.org/10.29056/jncist.2020.03.04
Abstract : Among the phases of self-driving, the transition of control over drivingvehicles to drivers occurs in the conditional self-driving phase. Therefore, designing appropriate user interfaces to perform safe control transitions is crucial. However, most previous studies have focused on evaluating the performance and safety of self-driving systems. In the conditional self-driving phase, no regulations that can force drivers to control their eyes currently exist, meaning the control transition, which relies solely on the visual-centric user interfaces used in traditional driving environments, is not safe. Therefore, in this study, we developed guidelines for an auditory user interface design to ensure safe control transition. We derived these guidelines by applying a user experience design methodology and defining important principles in the appropriate combination of voice and sound feedback.
Keyword : control transitions, auditory user interface, take-over, hand-over, conditional self-driving
http://dx.doi.org/10.29056/jncist.2020.03.05
Interactive media contents suggestion based on Geumgang painting conceptJeong seon
Abstract : With the development of the media content industry, the development of visual media is transforming the way all information is communicated into visual images. In recent years, visual literacy has been highlighted as an important element of cultural education, and the need for art education has been emphasized. As the interest in education of today's society meets IT technology, various contents such as video, AR, VR, hologram, media facade are created. In this paper, we selected these pictorial elements as the subject of contents and propose the development of interactive media contents of Geumgang painting, which incorporates projection mapping technology based on Jeong seon. Through this content of Geumgangsando, historical background knowledge can be learned based on the cultural aspects of Joseon Dynasty for various paintings in the video.
Keyword : Geumgangsando, Virtual Reality, Interactive Media Contents, Media facade, Hologram
http://dx.doi.org/10.29056/jncist.2020.03.06
The Relationship between Corporate Governance and Firm Value -A Comparative Study-
Abstract : This paper mainly discuss about that explain if there exists any relationship between corporate governance and firm value. Based on the prior research that corporate governance would play a role reducing the agency cost producing by the separation of ownership, this study to verify that corporate governance would reduce the firms’ financing cost. Specifically, the firms with high corporate governance are predicted to have higher profitability and firm value measured by Tobin’s Q. This study examined theses hypotheses, using a sample of 5,785 on Korean firms over ten-year period (2004-2013) and a number of 7,777 for the Chinese sample data over. An index published by Korean Corporate Governance Service (KCGS) and a RKS of China were used as the measure of corporate governance. The results of this paper find that the good corporate governance enhance the firm value measured by Tobin’s Q.
Keyword : Corporate Governance, Firm Value, Tobin’s Q
http://dx.doi.org/10.29056/jncist.2020.03.07
A Study of Strengthening Response Capabilities for Disaster Type and Analyzing Collaboration Systems
Abstract : In recent years, it has become increasingly difficult to predict whenhow a disaster will occur. The disaster response system can be said to be a cooperative structure of government-enterprise-citizen society based on the type of disaster and on-the-spot situation to reduce risk factors and minimize the spread of damages caused by a disaster. In order to operate an effective cooperative structure, a disaster response plan and a cooperative system plan are needed to define roles and responsibilities between governments, public institutions, NGOs, and private organizations in the event of a disaster. In this paper, for effective disaster response, the nature and characteristics of the disaster response organization and the main tasks of the disaster management support organization are summarized by disaster type. In addition, by establishing a unified command system for quick application to disaster response tasks and establishing the basic system of 13 collaboration functions for each disaster type for smooth collaboration and information sharing, the schematic content of the collaboration response system according to the size and complexity of disasters.
Keyword : Response rules, collaboration system, standard operation procedure, complex disaster
http://dx.doi.org/10.29056/jncist.2020.03.08
Abstract : Shoulder-related disorders include joint and labial injuries, shoulder impingement syndrome, habitual shoulder dislocation and slab lesions. According to health insurance statistics, the number of patients treated with these shoulder lesions is increasing every year. MRI (Magnetic Resonance Imaging) is a representative method of diagnosing such shoulder-related diseases. In this paper, we will use U-net, a convolution neural network algorithm, to automatically segment humeral heads, joints in MR images to help doctors diagnose related diseases more quickly. A total of 2251 MR images of 81 patients were collected and used as training set and test set. As a result of evaluation of U-net trained scapular segmentation model validated by chapter 427 data, sensitivity of 89.32%, specificity of 99.78%, accuracy of 99.44%, and die coefficient of 92.02%. In the future, image processing techniques will solve the ambiguity of the scapular border and collect additional data to fully train the joints and regions, which will provide a higher level of accuracy and help the diagnosis for doctors.
Keyword : Artificial Intelligence, Deep learning, Segmentation. Convolution Neural Network, Shoulder disease
http://dx.doi.org/10.29056/jncist.2020.03.09
Table of Contents