• The Journal of Korean Academy of Prosthodontics
• /
• v.50 no.3
• /
• pp.162-168
• /
• 2012

Purpose: Teeth are generally exposed when people smiling. Moreover, the exposed teeth and soft tissue when smiling becomes an important guideline for esthetically prosthetic restoration. This research is to compare and find out differences of Korean young men's smile living in presence and twenty years ago. Materials and methods: Subjects, 100 young men (50 male and 50 female), were required about several aspects; normally developed physical condition, no psychological or genetic disorders, a fine face with no loss of teeth, no experience in orthodontic or prosthetic treatment, relatively normal occlusion, aged between 20-29. The photos of the subject at rest position and front face when fully smiled were taken three times. 100 photos (50 male and 50 female) were chosen at random from the 240 university students' smile photos taken by Yoon and his colleagues in 1991. By Hulsey's method of measuring smile, several factors; the change of upper lip curvature, the change of the relation between the upper lip and teeth, parallelism between Mx. incisor and lower lip, contact relation between Mx. incisor and lower lip and teeth displayed in a smile, were measured and analyzed. Meanwhile, ten dentists assessed aesthetic evaluation about men and women's smile for twice and recorded and compared smile score. The -test (P<.05) was used to compare the measured value. The difference of smile score was analyzed by t-test (P<.05). Results: The smile score calculated in 2011 (60.22) was higher than that of 1991 (52.80). Among five measurement categories, the noticeable difference was distinguished from two factors; the change of upper lip curvature and contact relation between Mx. incisor and lower lip. Conclusion: The Korean young men's smile has been considerably improved for twenty years. And it is found that the change of upper lip curvature plays an important role, that is, the smile formed with an ascended labial commissure has been increased significantly.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.175-197
• /
• 2021

Recently, with the development of deep learning technology, research on unstructured data analysis is being actively conducted, and it is showing remarkable results in various fields such as classification, summary, and generation. Among various text analysis fields, text classification is the most widely used technology in academia and industry. Text classification includes binary class classification with one label among two classes, multi-class classification with one label among several classes, and multi-label classification with multiple labels among several classes. In particular, multi-label classification requires a different training method from binary class classification and multi-class classification because of the characteristic of having multiple labels. In addition, since the number of labels to be predicted increases as the number of labels and classes increases, there is a limitation in that performance improvement is difficult due to an increase in prediction difficulty. To overcome these limitations, (i) compressing the initially given high-dimensional label space into a low-dimensional latent label space, (ii) after performing training to predict the compressed label, (iii) restoring the predicted label to the high-dimensional original label space, research on label embedding is being actively conducted. Typical label embedding techniques include Principal Label Space Transformation (PLST), Multi-Label Classification via Boolean Matrix Decomposition (MLC-BMaD), and Bayesian Multi-Label Compressed Sensing (BML-CS). However, since these techniques consider only the linear relationship between labels or compress the labels by random transformation, it is difficult to understand the non-linear relationship between labels, so there is a limitation in that it is not possible to create a latent label space sufficiently containing the information of the original label. Recently, there have been increasing attempts to improve performance by applying deep learning technology to label embedding. Label embedding using an autoencoder, a deep learning model that is effective for data compression and restoration, is representative. However, the traditional autoencoder-based label embedding has a limitation in that a large amount of information loss occurs when compressing a high-dimensional label space having a myriad of classes into a low-dimensional latent label space. This can be found in the gradient loss problem that occurs in the backpropagation process of learning. To solve this problem, skip connection was devised, and by adding the input of the layer to the output to prevent gradient loss during backpropagation, efficient learning is possible even when the layer is deep. Skip connection is mainly used for image feature extraction in convolutional neural networks, but studies using skip connection in autoencoder or label embedding process are still lacking. Therefore, in this study, we propose an autoencoder-based label embedding methodology in which skip connections are added to each of the encoder and decoder to form a low-dimensional latent label space that reflects the information of the high-dimensional label space well. In addition, the proposed methodology was applied to actual paper keywords to derive the high-dimensional keyword label space and the low-dimensional latent label space. Using this, we conducted an experiment to predict the compressed keyword vector existing in the latent label space from the paper abstract and to evaluate the multi-label classification by restoring the predicted keyword vector back to the original label space. As a result, the accuracy, precision, recall, and F1 score used as performance indicators showed far superior performance in multi-label classification based on the proposed methodology compared to traditional multi-label classification methods. This can be seen that the low-dimensional latent label space derived through the proposed methodology well reflected the information of the high-dimensional label space, which ultimately led to the improvement of the performance of the multi-label classification itself. In addition, the utility of the proposed methodology was identified by comparing the performance of the proposed methodology according to the domain characteristics and the number of dimensions of the latent label space.

• Maritime Security
• /
• v.1 no.1
• /
• pp.215-240
• /
• 2020

In 2017, the U.S. DARPA coined 'mosaic warfare' as a new way of warfighting. According to the Timothy Grayson, director of DARPA's Strategic Technologies Office, mosaic warfare is a "system of system" approach to warfghting designed around compatible "tiles" of capabilities, rather than uniquely shaped "puzzle pieces" that must be fitted into a specific slot in a battle plan in order for it to work. Prior to cover mosaic warfare theory and recent development, it deals analyze its background and several premises for better understanding. The U.S. DoD officials might acknowledge the current its forces vulnerability to the China's A2/AD assets. Furthermore, the U.S. seeks to complete military superiority even in other nation's territorial domains including sea and air. Given its rapid combat restoration capability and less manpower casualty, the U.S. would be able to ready to endure war of attrition that requires massive resources. The core concept of mosaic warfare is a "decision centric warfare". To embody this idea, it create adaptability for U.S. forces and complexity or uncertainty for the enemy through the rapid composition and recomposition of a more disag g reg ated U.S. military force using human command and machine control. This allows providing more options to friendly forces and collapse adversary's OODA loop eventually. Adaptable kill web, composable force packages, A.I., and context-centric C3 architecture are crucial elements to implement and carry out mosaic warfare. Recently, CSBA showed an compelling assessment of mosaic warfare simulation. In this wargame, there was a significant differences between traditional and mosaic teams. Mosaic team was able to mount more simultaneous actions, creating additional complexity to adversaries and overwhelming their decision-making with less friendly force's human casualty. It increase the speed of the U.S. force's decision-making, enabling commanders to better employ tempo. Consequently, this article finds out and suggests implications for Korea armed forces. First of all, it needs to examine and develop 'mosaic warfare' in terms of our security circumstance. In response to future warfare, reviewing overall force structure and architecture is required which is able to compose force element regardless domain. In regards to insufficient defense resources and budget, "choice" and "concentration" are also essential. It needs to have eyes on the neighboring countries' development of future war concept carefully.

• Journal of Korea Water Resources Association
• /
• v.56 no.3
• /
• pp.195-210
• /
• 2023

Additional deposition and erosion in large rivers in South Korea have continued to occur toward morphological stabilization after massive dredging through the four major river restoration project, subsequently requiring precise bathymetry monitoring. Hyperspectral bathymetry method has increasingly been highlighted as an alternative way to estimate bathymetry with high spatial resolution in shallow depth for replacing classical intrusive direct measurement techniques. This study introduced the conventional Optimal Band Ratio Analysis (OBRA) of hyperspectral bathymetry method, and evaluated the performance in a domestic large river in normal turbid and flow condition. Maximum measurable depth was estimated by applying correlation coefficient and root mean square error (RMSE) produced during OBRA with cascadedly applying cut-off depth, where the consequent hyperspectral bathymetry map excluded the region over the derived maximum measurable depth. Also non-linearity was considered in building relation between optimal band and depth. We applied the method to the Nakdong and Hwang River confluence as a large river case and obtained the following features. First, the hyperspectal method showed acceptable performance in morphological mapping for shallow regions, where the maximum measurable depth was 2.5 m and 1.25 m in the Nakdong and Hwang river, respectively. Second, RMSE was more feasible to derive the maximum measurable depth rather than the conventional correlation coefficient whereby considering various scenario of excluding range of in situ depths for OBRA. Third, highly turbid region in Hwang River did not allow hyperspectral bathymetry mapping compared with the case of adjacent Nakdong River, where maximum measurable depth was down to half in Hwang River.

• Horticultural Science & Technology
• /
• v.34 no.3
• /
• pp.461-471
• /
• 2016

This study was conducted to establish an in vitro propagation system for sea-milkwort (Glaux maritima L.), which is an endangered coastal plant species with high horticultural value. Two phenotypes, 'Red type (RT)' and 'Pistachio type (PT)' based on the colors of stem and flower, were obtained from a personal horticulturist in 2009 and used for this study as plant materials. The stock plants showed typical morphologies in flower, capsule, and seed appearances as previously reported. Low temperature treatment at $4^{\circ}C$ for four or more weeks after in vitro sowing maximized seed germination percentage, indicating that imbibition of seed and subsequent low temperature treatment are crucial for its germination. The in vitro seedlings had phenotypic variation, falling into 'RT' and 'PT' classes like the stock plants. Although slight differences depending on genotype and medium were recognized, the fourth or fifth nodes detached from the in vitro seedlings revealed the best multiplication efficacy when estimated on the basis of total number of nodes of newly developed axillary shoots. In addition, the nodes from 'RT' and 'PT' regenerated the most shoots on medium supplemented with $0.5mg{\cdot}L^{-1}$ BA alone and $0.5mg{\cdot}L^{-1}$ BA plus $0.5mg{\cdot}L^{-1}$ IAA, respectively. The node culture-derived plantlets were well acclimatized in a culture room ex vitro and completed the pseudo-annual life cycle coincident with that in the natural salt march habitat with the current cultivation method of applying fresh water-irrigation under an inland environment. This work represents the first report of in vitro propagation of sea-milkwort. Thus, our study will contribute to exo-habitat conservation and natural habitat restoration of this endangered species in addition to development of a horticultural product.

• Korean Journal of Plant Resources
• /
• v.31 no.4
• /
• pp.384-395
• /
• 2018

This study was carried out to investigate the environment factors including community structure and soil characteristics in the wild habitats of Cephalotaxus koreana, and offers the basic information for habitats conservation and restoration. Most of the wild habitats were located at altitudes between 148~835 m with inclinations ranged as $12{\sim}32^{\circ}$. The average soil pH was 4.7~5.9, soil organic matter was 5.72~15.99%, cation exchange capacity was $14.1{\sim}19.9cmolc/kg^{-1}$ and exchangeable $K^+$, $Ca^{2+}$, $Mg^{2+}$ was 0.25~0.48 cmolc/kg, 0.79~6.68 cmolc/kg, 0.31~1.73 cmolc/kg, respectively. The dominant species of tree layer were found to be dominated by Quercus dentata in Jekbo-san (C1), Acer pictum in Bogae-san (C2), Acer pseudosieboldianum in Geumwon-san (C3), Q. serrata in Jiri-san (C4), Zelkova serrata in Baegun-san (C5), and Q. acutissima in Duryun-san (C6). The Species diversity (H') was 0.854~1.234, evenness (J') was 0.654~0.993, and dominance (D) was found to be 0.067~0.346. Correlation coefficients analysis based on environmental factors, community structure and value of species diversity shows that growth of Cephalotaxus koreana is correlated with species diversity and evenness. This result show that Cephalotaxus koreana habitats located in mature stands.