• Korean Journal of Heritage: History & Science
• /
• v.43 no.3
• /
• pp.306-333
• /
• 2010

Previous studies on Wolseong fortress have focused on capital system of Silla Dynasty and on the recreation of Wolseong fortress due to the excavations in and around Wolseong moat. Since the report on the Geographical Survey of Wolseong fortress was published and GPR survey in Wolseong fortress was executed as a trial test in 2004, the academic interest in the site has now expanded to the inside of the fortress. From such context, the preliminary research on the fortress including geophysical survey had been commenced. GPR survey had been conducted for a year from March, 2007. The principal purpose of the recent 3D GPR survey was to provide visualization of subsurface images of the entire Wolseong fortress area. In order to obtain 3D GPR data, dense profile lines were laid in grid-form. The total area surveyed was $112,535m^2$. Depth slice was applied to analyse each level to examine how the layers of the remains had changed and overlapped over time. In addition, slice overlay analysis methodology was used to gather reflects of each depth on a single map. Isolated surface visualization, which is one of 3D analysis methods, was also employed to gain more in-depth understanding and more accurate interpretations of the remain The GPR survey has confirmed that there are building sites whose archaeological features can be classified into 14 different groups. Three interesting areas with huge public building arrangement have been found in Zone 2 in the far west, Zone 9 in the middle, and Zone 14 in the far east. It is recognized that such areas must had been used for important public functions. This research has displayed that 3D GPR survey can be effective for a vast area of archaeological remains and that slice overlay images can provide clearer image with high contrast for objects and remains buried the site.

• Korean Journal of Heritage: History & Science
• /
• v.43 no.2
• /
• pp.60-81
• /
• 2010

Jeju Island, in Korea, shows many characteristics that are differentiated from the rest of Korea. Its culture is rooted in mythology which advocates a egalitarian, rather than hierarchical, social structure, the place of women in the home is relatively high, and the formation of buildings, the separation of cooking and heating facilities, and the living format of residential homes is dissimilar. These disparities in culture indicate that Jeju Island's heritage was not formed only from influences from the North, but also from other places as well. To fill in the blanks, residential homes in Jeju Island were compared with those scattered throughout the East China Sea, which connect the southern coastline of the Korean peninsula and Jeju Island. The regions encompassed by the East China Sea, sharing the Kuroshio current and a seasonal wind, can be considered as one cultural region integrating cultural aspects from the continental North and the oceanbound South. The unique characteristics of southern culture as seen in southern residences was examined through an investigation of the sacred places in which gods were considered to dwell. First, the myths of these areas usually concerned with the ocean, and a sterile environment made sustenance impossible without a dual livelihood, usually taking on the forms of half-farming and half-fishing, or half-farming, half-gardening. Although family compositions were strongly matricentric or collateral thanks to southern influence, a patriarchical system like those found in the North were present in the upper classes and in the cities. Therefore, residential spaces were not divided based on age or gender, as in hierarchical societies, but according to family and function. Second, these areas had local belief systems based on animism and ancestor worship, and household deities were closely related to women, agriculture and fire. The deities of the kitchen, the granary and the toilet were mostly female, and the role of priest was often filled by a woman. After Buddhism and Confucianism were introduced from mainland Korea, China and Japan, the sacred areas of the household took on a dual form, integrating the female-focused local rites with male-centered Buddhist and Confucian rites. Third, in accordance with worship of a kitchen deity, a granary deity, and a toilet deity led to these areas of the home being separated into disparate buildings. Eventually, these areas became absorbed into the home as architectural technology was further developed and lifestyles were changed. There was also integration of northern and southern cultures, with rites concerning granary and toilet deities coming from China, and the personality of the kitchen deity being related to the southern sea. In addition, the use of stone in separate kitchens, granaries, and toilets is a distinguishing characteristic of the East China Sea. This research is a part of the results gained from a project funded by the Korea Research Foundation in 2006.

• Korean Journal of Heritage: History & Science
• /
• v.42 no.4
• /
• pp.198-221
• /
• 2009

Royal tombs of Joseon dynasty are the crystallization of history, ideology, culture, art, architecture, and ritual ceremony of Joseon dynasty, all mingled in one. So, they are very significant symbols showing 500 years of dynastic history as a whole. Among various factors comprising a royal tomb, stone figures surrounding grave mound are special factors as a symbol protecting it. Further among them, twelve zodiac images arrayed nearest to the grave mound represent the core of the function. Images of twelve zodiac signs originated from the tombs of the Unified Silla Kingdom are certain to hold important role and position in the construction of royal tombs, judging from huge scale and excellent sculptural art of them. However, both their scale and form had been gradually simplified in Goryeo and Joseon dynasty, thus the importance of them has been underestimated compared to other stone figures Images of twelve zodiac signs were very important factors which decorated royal tombs both as a protective role and as a concept of direction. Their historicity and symbolism cannot be neglected in that they had been transmitted to the royal tombs of Joseon dynasty. In this paper, images of twelve zodiac signs expressed in the royal tombs of Joseon dynasty are classified into 3 forms, and reviewed the origin and development of them for each period. They could be classified into 3 forms ; civil vassals with human body and head, civil vassals with human body and head wearing hat decorated with zodiac animals, Chinese characters of either zodiac signs or either a combination of 10 calendar signs and 8 trigrams. The above 3 forms originated from China and became a favorite motif to decorate the royal tombs from early Joseon period until late Joseon by replacing each other and thus changing along the course of the dynasty. In the meantime, we can see a unique character in the images of twelve zodiac signs of royal tombs of Joseon dynasty. In some cases, 24 directions are expressed in which 10 calender signs and 8 trigrams are composed altogether. Images of twelve zodiac signs in the royal bombs of Joseon dynasty are very significant as evidences by which we can confirm uniqueness and tradition of Korean tomb system transmitted from Unified Silla period.

• Journal of Intelligence and Information Systems
• /
• v.25 no.3
• /
• pp.43-62
• /
• 2019

At one time, the anomaly detection sector dominated the method of determining whether there was an abnormality based on the statistics derived from specific data. This methodology was possible because the dimension of the data was simple in the past, so the classical statistical method could work effectively. However, as the characteristics of data have changed complexly in the era of big data, it has become more difficult to accurately analyze and predict the data that occurs throughout the industry in the conventional way. Therefore, SVM and Decision Tree based supervised learning algorithms were used. However, there is peculiarity that supervised learning based model can only accurately predict the test data, when the number of classes is equal to the number of normal classes and most of the data generated in the industry has unbalanced data class. Therefore, the predicted results are not always valid when supervised learning model is applied. In order to overcome these drawbacks, many studies now use the unsupervised learning-based model that is not influenced by class distribution, such as autoencoder or generative adversarial networks. In this paper, we propose a method to detect anomalies using generative adversarial networks. AnoGAN, introduced in the study of Thomas et al (2017), is a classification model that performs abnormal detection of medical images. It was composed of a Convolution Neural Net and was used in the field of detection. On the other hand, sequencing data abnormality detection using generative adversarial network is a lack of research papers compared to image data. Of course, in Li et al (2018), a study by Li et al (LSTM), a type of recurrent neural network, has proposed a model to classify the abnormities of numerical sequence data, but it has not been used for categorical sequence data, as well as feature matching method applied by salans et al.(2016). So it suggests that there are a number of studies to be tried on in the ideal classification of sequence data through a generative adversarial Network. In order to learn the sequence data, the structure of the generative adversarial networks is composed of LSTM, and the 2 stacked-LSTM of the generator is composed of 32-dim hidden unit layers and 64-dim hidden unit layers. The LSTM of the discriminator consists of 64-dim hidden unit layer were used. In the process of deriving abnormal scores from existing paper of Anomaly Detection for Sequence data, entropy values of probability of actual data are used in the process of deriving abnormal scores. but in this paper, as mentioned earlier, abnormal scores have been derived by using feature matching techniques. In addition, the process of optimizing latent variables was designed with LSTM to improve model performance. The modified form of generative adversarial model was more accurate in all experiments than the autoencoder in terms of precision and was approximately 7% higher in accuracy. In terms of Robustness, Generative adversarial networks also performed better than autoencoder. Because generative adversarial networks can learn data distribution from real categorical sequence data, Unaffected by a single normal data. But autoencoder is not. Result of Robustness test showed that he accuracy of the autocoder was 92%, the accuracy of the hostile neural network was 96%, and in terms of sensitivity, the autocoder was 40% and the hostile neural network was 51%. In this paper, experiments have also been conducted to show how much performance changes due to differences in the optimization structure of potential variables. As a result, the level of 1% was improved in terms of sensitivity. These results suggest that it presented a new perspective on optimizing latent variable that were relatively insignificant.

• Korean Journal of Heritage: History & Science
• /
• v.47 no.4
• /
• pp.22-47
• /
• 2014

This study was to understand the basic principles of the East Asian wooden structure system research and analysis. The Korea and China ancient architecture internal structure research that the combination of girders and crossbeams position. The ancient wooden structures of eastern Asian countries, Korea and China are not much different from each other in the principles of the wooden architecture structure, combining pillars, purlins and crossbeams. However, it seems that age-division, local-division, national-division differs in detail techniques. China ancient wooden structures combination of purlin and crossbeam, and So-seul Timber(Chinese name: Chashou叉手, Tuojiao 托脚) seems to show differences according to the age of the fulcrum position, detailed approach is also different according to various historical dynasty. Before in the 15th century, Purlin and Crossbeam are coupled to each other, but since the 15th century, seems to have developed a technique combined with each other Girder and Crossbeam and to prevent buckling of the Crossbeam cross-sectional area increased dramatically. For Tuojiao in China Tang-Wudai dynasty(A.D. 618~979), can see that saw the top position Girder and Tuojiao no direct coupling, can be seen as maintaining the safety of the material than the material of the inner wooden structures prevent buckling of the purlin. Korea ancient wooden structures of Goryeo dynasty(A.D. 918~1391), So-seul Timber(Chinese name Tuojiao) why do not to use the fashion? To use Purlin Lower backing material techniques to prevent buckling is a popular trend to stable can be thought of as a preferred way to maintain. I think that with universality beyond the local-division, national-division and the two countries since the 15th century of Korea and China ancient wooden structures detailed mechanism for the purlin buckling. In middle-late Chosen dynasty, The effect of Deotgeolyi- techniques and fleeting beams reduce the purlin buckling that reduces the load transmitted from purlin and crossbeam of how to reduce the load on the roof portion of the architecture fleeting beams used, which of craftsmanship of the Chosen Dynasty building can be referred to as another technique for preventing buckling purlin. This Korea and China ancient architecture purlin beam structure and material So-seul Timber study. Seems to be able to provide a basic research study to restore and designed the old wooden architectures.

• Journal of Korean Society of Forest Science
• /
• v.108 no.3
• /
• pp.290-295
• /
• 2019

This study was carried out to investigate the possibility of establishing a reproductive system for the seed of Pedicularis hallaisanensis, which is in the endangered wild species II class in Korea. The seed of P. hallaisanensis is egg-shaped, and the seed coat is dark brown. The embryo was identified as a dwarf type by the seed section. The seed length was $0.47{\pm}0.07mm$, width $0.16{\pm}0.006mm$, and thickness $0.12{\pm}0.01mm$. The weight of one seed was $0.0003{\pm}0.0001mg$, and 1000 seeds weighed $4.59{\pm}0.02mg$. The degree of seed viability was 75.33% by the tetrazolium (TZ) assay. The highest germination rate of P. hallaisanensis seed was 71% after 4 weeks of storage at $4^{\circ}C$. However, the germination rate tended to decrease gradually over a longer storage period. The germination rates after 6 or 8 weeks of storage at $4^{\circ}C$ were 64% and 60%, respectively. We used two host plants, Artemisia princeps and Dendranthema zawadskii, to determine the effect of host plants on P. hallaisanensis seed germination. The germination of P. hallaisanensis mixed with A. princeps or D. zawadskii started at 53.5 and 62.5 days after sowing, respectively. We did not find any germination 164 days postsowing with both host plants. When A. princeps and D. zawadskii were used as host plants for P. hallaisanensis seed germination, P. hallaisanensis seed germination rates were 45.5% and 19.5%, respectively. The average time to germination was 70.2 days for A. princeps, and 46.8 days for D. zawadskii.

• Journal of Intelligence and Information Systems
• /
• v.25 no.3
• /
• pp.201-220
• /
• 2019

Fake news has emerged as a significant issue over the last few years, igniting discussions and research on how to solve this problem. In particular, studies on automated fact-checking and fake news detection using artificial intelligence and text analysis techniques have drawn attention. Fake news detection research entails a form of document classification; thus, document classification techniques have been widely used in this type of research. However, document summarization techniques have been inconspicuous in this field. At the same time, automatic news summarization services have become popular, and a recent study found that the use of news summarized through abstractive summarization has strengthened the predictive performance of fake news detection models. Therefore, the need to study the integration of document summarization technology in the domestic news data environment has become evident. In order to examine the effect of extractive summarization on the fake news detection model, we first summarized news articles through extractive summarization. Second, we created a summarized news-based detection model. Finally, we compared our model with the full-text-based detection model. The study found that BPN(Back Propagation Neural Network) and SVM(Support Vector Machine) did not exhibit a large difference in performance; however, for DT(Decision Tree), the full-text-based model demonstrated a somewhat better performance. In the case of LR(Logistic Regression), our model exhibited the superior performance. Nonetheless, the results did not show a statistically significant difference between our model and the full-text-based model. Therefore, when the summary is applied, at least the core information of the fake news is preserved, and the LR-based model can confirm the possibility of performance improvement. This study features an experimental application of extractive summarization in fake news detection research by employing various machine-learning algorithms. The study's limitations are, essentially, the relatively small amount of data and the lack of comparison between various summarization technologies. Therefore, an in-depth analysis that applies various analytical techniques to a larger data volume would be helpful in the future.

• Korean Journal of Heritage: History & Science
• /
• v.36
• /
• pp.347-377
• /
• 2003

China is a country with large agricultural areas and subject to frequent calamities. Drought is the top of them. It has been a key problem for development of agriculture in the country. In the long struggle against drought, Chinese have accumulated many rational and irrational experiences. The Dragon Kings Belief, which is popular in North China and discussed in a thesis, is one of their irrational experiences. The belief was passed together with Buddhism from India to China in the Tang Dynasty. After it settled down, it was incorporated with the local five dragons belief and a set of beliefs in dragon kings came into existence. The emergence of the dragon kings belief ended the history that the title of rain got was not clear in China and Dragon kings finally got the status. Irrigation is the lifeblood of agriculture in China. In a Chinese mind, Dragon kings are the most important gods who take charge of rain and thus offer the lifeblood. In understanding the nature and characteristics of Chinese traditional culture, it is important for us to make clear the origin and evolution of the belief, find out its nature, function and operation. In the every year beginning of February of the Fanzhuang calendar in the people of Hebeisheng Zhaoxian, would all hold a festival to offer sacrifices to the $^{{\circ}TM}^{\prime}longpai$. Longpai was regarded as the core of the temple fair, thus the native sons came to call this festival; "longpaihui". In this region the'Fanzhuang longpaihui'developed into a well knownand grand temple fair. It was able to attract numerous pilgrims with its special magic power, occupying a place in $China^{{\circ}TM}$ 'eryueer'festival with festive dragon activities. The dragon is a common totem among Chinese nationals. The belief worship of the dragon dates from the start time of primitive societies. Dragon oneself the ancients worship's thunder lightning. In the worship of the great universe, at first afterwards this belief with the tribe's totem worships to combine to become the animal spirit. In ancient myths legends, along with folk religion and beliefs all hold a very important position. The longpaihui is a temple fair without a temple; this characteristic is a distinction between longpaihui and other temple fairs. As for longpaihui must of the early historical records are unclear. The originator of a huitou system has a kind of organized form of the special features rather, originator of a huitou not fix constant, everything follows voluntarily principle, can become member with the freedom, also can back at any time the meeting. There is a longpaihui for 'dangjiaren', is total representative director in the originator of a huitou will. 'banghui' scope particularly for extensive, come apparently every kind of buildup that help can return into the banghui, where is the person of this village or outside village of, the general cent in banghui work is clear and definite, for longpaihui would various businesses open smoothly the exhibition provides to guarantees powerfully. Fanzhuang longpaihui from the beginning of February to beginning six proceed six days totally. The longpai is used as the ancestry absolute being to exsits with the community absolute being at the same time in fanzhuang first took civil faith, in reality is a kind of method to support social machine in native folks realize together that local community that important function, it provided a space, a kind of a view to take with a relation, rising contact, communication, solidify the community contents small village, formation with fanzhuang. The fanzhuang is used as supplies for gathering town, by luck too is this local community trade exchanges center at the same time therefore can say the faith of the longpai, in addition to its people's custom, religious meaning, still have got the important and social function. Moreover matter worthy of mentioning, Longpai would in organize process, from prepare and plan the producing of meeting every kind of meeting a longpeng of the matter do, all letting person feeling is to adjust the popular support of, get the mass approbation with positive participate. Apart from the originator of a huitou excluding, those although not originator of a huitou, however enthusiasm participate the banghui of its business, also is too much for the number.

• Economic and Environmental Geology
• /
• v.54 no.2
• /
• pp.177-186
• /
• 2021

For disaster management and mitigation of earthquakes in Korea Peninsula, active fault investigation has been conducted for the past 5 years. In particular, investigation of sediment-covered active faults integrates geomorphological analysis on airborne LiDAR data, surface geological survey, and geophysical exploration, and unearths subsurface active faults by trench survey. However, the fault traces revealed by trench surveys are only available for investigation during a limited time and restored to the previous condition. Thus, the geological data describing the fault trench sites remain as the qualitative data in terms of research articles and reports. To extend the limitations due to temporal nature of geological studies, we utilized a terrestrial LiDAR to produce 3D point clouds for the fault trench sites and restored them in a digital space. The terrestrial LiDAR scanning was conducted at two trench sites located near the Yangsan Fault and acquired amplitude and reflectance from the surveyed area as well as color information by combining photogrammetry with the LiDAR system. The scanned data were merged to form the 3D point clouds having the average geometric error of 0.003 m, which exhibited the sufficient accuracy to restore the details of the surveyed trench sites. However, we found more post-processing on the scanned data would be necessary because the amplitudes and reflectances of the point clouds varied depending on the scan positions and the colors of the trench surfaces were captured differently depending on the light exposures available at the time. Such point clouds are pretty large in size and visualized through a limited set of softwares, which limits data sharing among researchers. As an alternative, we suggested Potree, an open-source web-based platform, to visualize the point clouds of the trench sites. In this study, as a result, we identified that terrestrial LiDAR data can be practical to increase reproducibility of geological field studies and easily accessible by researchers and students in Earth Sciences.

• 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.