• Economic and Environmental Geology
• /
• v.40 no.1 s.182
• /
• pp.115-128
• /
• 2007

This study proposed a modified logistic regression model for a probabilistic prediction of debris flow on natural terrain at the granitic rock area. The modified model dose not contain any categorical factors that were used in the previous model and secured higher reliability of prediction than that of the previous one. The modified model is composed of lithology, two factors of geomorphology, and three factors of soil property. Verification result shows that the prediction reliability is more than 86%. Using the modified regression model, the landslide prediction maps were established. In case of Sacheon area, the prediction map showed that the landslide occurrence was not well corresponded with the model since, even though the forest-fred area was distributed on the center of the model, no factors were considered for the landslide predictions. On the other hand, the prediction model was well corresponded with landslide occurrence at Jumunjin-Yeongok area. The prediction model developed in this study has very high availability to employ in other granitic areas.

• Tunnel and Underground Space
• /
• v.30 no.5
• /
• pp.462-472
• /
• 2020

Rock classification is fundamental discipline of exploring geological and geotechnical features in a site, which, however, may not be easy works because of high diversity of rock shape and color according to its origin, geological history and so on. With the great success of convolutional neural networks (CNN) in many different image-based classification tasks, there has been increasing interest in taking advantage of CNN to classify geological material. In this study, a feasibility of the deep CNN is investigated for automatically and accurately identifying rock types, focusing on the condition of various shapes and colors even in the same rock type. It can be further developed to a mobile application for assisting geologist in classifying rocks in fieldwork. The structure of CNN model used in this study is based on a deep residual neural network (ResNet), which is an ultra-deep CNN using in object detection and classification. The proposed CNN was trained on 10 typical rock types with an overall accuracy of 84% on the test set. The result demonstrates that the proposed approach is not only able to classify rock type using images, but also represents an improvement as taking highly diverse rock image dataset as input.

• Journal of the Korean Geotechnical Society
• /
• v.20 no.9
• /
• pp.25-32
• /
• 2004

Some amount of cements can be added into the soil with high water content to improve the engineering properties. In such a case, it is difficult to predict and figure out the phase changes of the soil-cement mixture which is closely associated with workability of the soil-cement mixture. Changes in heat of hydration and hardness of the cement pastes are known to provide the useful information about the phase changes of the soil-cement mixtures. In this study, heat of hydration and cone penetration depth were measured from the specimens of cement paste and 3 soil-cement mixtures. From the experimental results, it was found that the phase changes of the soil-cement mixtures are the same as those of cement paste, and that shear strength of the mixtures abruptly increases when the heat of hydration is minimum. Initial setting time of the mixtures coincides with the state when fall cone penetration depth was 1.0 mm and it is defined as plastic limit of the mixtures. Initial setting time of the mixtures is retarded as soil/cement ratio is increased. Measurements of heat of hydration and fall cone apparatus could be the useful tools to predict the phase changes of tile soil-cement mixtures.

• Journal of agriculture & life science
• /
• v.46 no.5
• /
• pp.1-10
• /
• 2012

This study analyzed on the characteristics of vegetation structure, species composition and DBH class distribution in order to conservation and effective management for Taxus cuspidata community in Mt. Seorak, Mt. Balwang, Mt. Taebaek, and Mt. Odae. The vegetation in upper, subtree and shrub layer was consist of 11, 22, 33 species in Mt. Seorak, 15, 21, 33 species in Mt. Balwang, 10, 23, 36 species in Mt. Taebaek, and 14, 30, 32 species in Mt. Odae. As a result of importance value at all study sites, T. cuspidata and Abies nephrolepis in upper layer, T. cuspidata, A. nephrolepis and Acer komarovii in subtree layer, and Tripterygium regelii in shrub layer were high, respectively. Species diversity in upper and subtree layer at all study sited were ranged 0.834~1.234 and 1.125~1.329, respectively. According to the DBH class of major three species, T. cuspidata in Mt. Odae site showed a reverse J-shaped curve, which was estimated that T. cuspidata community of this site might be maintained continuously as a stable state.

• Journal of agriculture & life science
• /
• v.46 no.6
• /
• pp.33-41
• /
• 2012

This study was conducted to provide the informations for conservation and effective management of Picea yezoensis community in Mt. Deogyu and Mt. Gyebang. The vegetation of tree, subtree and shrub layer was consist of 8, 20, 26 species in Mt. Deogyu, and 12, 23, 33 species in Mt. Gyebang. Importance value by layer P. yezoensis, Betula ermanii, Abies koreana at tree layer, B. ermanii, Quercus mongolica at subtree layer, and Sasa borealis at shrub layer in Mt. Deogyu, and P. yezoensis, B. ermanii, Abies nephrolepis at tree layer, Acer komarovii and A. ukurunduense at subtree layer, and Tripterygium regelii at shrub layer in Mt. Gyebang were high, respectively. Species diversity in Mt. Deogyu and Mt. Gyebang were 0.779 and 0.984 at tree layer, 1.052 and 1.161 at subtree layer, and 0.823 and 1.304 at shrub layer, respectively. According to the DBH class of major species, P. yezoensis in Mt. Deogyu showed a reverse J-shaped curve, which was estimated that P. yezoensis community of this site might be maintained continuously as a stable state.

• The Journal of Society for e-Business Studies
• /
• v.26 no.2
• /
• pp.19-43
• /
• 2021

The prediction of box office performance in performing arts institutions is an important issue in the performing arts industry and institutions. For this, traditional prediction methodology and data mining methodology using standardized data such as cast members, performance venues, and ticket prices have been proposed. However, although it is evident that audiences tend to seek out their intentions by the performance guide poster, few attempts were made to predict box office performance by analyzing poster images. Hence, the purpose of this study is to propose a deep learning application method that can predict box office success through performance-related poster images. Prediction was performed using deep learning algorithms such as Pure CNN, VGG-16, Inception-v3, and ResNet50 using poster images published on the KOPIS as learning data set. In addition, an ensemble with traditional regression analysis methodology was also attempted. As a result, it showed high discrimination performance exceeding 85% of box office prediction accuracy. This study is the first attempt to predict box office success using image data in the performing arts field, and the method proposed in this study can be applied to the areas of poster-based advertisements such as institutional promotions and corporate product advertisements.

• Asian-Australasian Journal of Animal Sciences
• /
• v.21 no.8
• /
• pp.1159-1163
• /
• 2008

Data of 42 balance measurements from dry and lactating Holstein cows and blood and urine samples from 24 Japanese Black cows were collected to evaluate the potential for predicting urinary nitrogen (N) excretion from plasma urea nitrogen (PUN). Similar positive correlations were obtained between N intake and apparent N absorption in dry and lactating cows. The regression equations of N intake on urinary N excretion varied in dry and lactating cows, and the difference of urinary N excretion between dry and lactating cows was due to the N secretion into milk. Highly positive correlations were observed between urinary N contents and urinary urea N in Japanese Black cows, and urinary urea N increased with increasing PUN. There were positive correlations between N intake and PUN in dry and lactating cows, but PUN and urinary N excretion in lactating cows were higher than in dry cows. There were positive correlations between PUN and urinary N excretion per BW in dry and lactating cows. Although urinary N excretion could be calculated as (N clearance rate of kidneys)PUNBW, high N clearance rate of kidneys, such as 2.08 L/d/kg BW, may be suitable to calculate urinary N excretion in lactating cows, compared with 1.33 L/d/kg BW in dry cows.

• Journal of Biomedical Engineering Research
• /
• v.41 no.2
• /
• pp.101-106
• /
• 2020

Although benign gastric ulcers do not develop into gastric cancer, they are similar to early gastric cancer and difficult to distinguish. This may lead to misconsider early gastric cancer as gastric ulcer while diagnosing. Since gastric cancer does not have any special symptoms until discovered, it is important to detect gastric ulcers by early gastroscopy to prevent the gastric cancer. Therefore, we developed a Convolution Neural Network (CNN) model that can be helpful for endoscopy. 3,015 images of gastroscopy of patients undergoing endoscopy at Gachon University Gil Hospital were used in this study. Using ResNet-50, three models were developed to classify normal and gastric ulcers, normal and gastric cancer, and gastric ulcer and gastric cancer. We applied the data augmentation technique to increase the number of training data and examined the effect on accuracy by varying the multiples. The accuracy of each model with the highest performance are as follows. The accuracy of normal and gastric ulcer classification model was 95.11% when the data were increased 15 times, the accuracy of normal and gastric cancer classification model was 98.28% when 15 times increased likewise, and 5 times increased data in gastric ulcer and gastric cancer classification model yielded 87.89%. We will collect additional specific shape of gastric ulcer and cancer data and will apply various image processing techniques for visual enhancement. Models that classify normal and lesion, which showed relatively high accuracy, will be re-learned through optimal parameter search.

• Journal of Plant Biotechnology
• /
• v.30 no.2
• /
• pp.179-184
• /
• 2003

In order to establish micropropagation system Anthurium andreanum 'Atlanta', dwarf type, shoots of A. andreanum were cultured on medium supplemented with cytokinin. Callus was formed from the base of shoots. high frequency callus induction was obtained on medium with 10.0mg/L BA or 10.0mg/L TDZ(thidiazuron) at more than 71.8%. The shoots were cultured on media with various combinations and concentrations of TDZ, BA and 2.4-D to enhance callus induction. Callus was induced at more than 72.6% and grew vigorously on media containing 10.0mg/L BA and 0.0∼0.5mg/L 2.4-D, or 1.0mg/L TDZ. Stimulation effects of cytokinin by 2.4-D did not occur in combined treatments of cytokinin and 2.4-D. Callus was cut into sections(7${\times}$10mm), and then cultured on media with BA alone or BA and 2.4-D to regenerate shoots and to stimulate the callus growth. Shoot regeneration and callus growth were effective on media with 10.0mg/L BA alone, or 10.0mg/L BA and 0.1mg/L 2.4-D. In combined treatments of BA and 2.4-D, stmulation effects of cytocinin by 2.4-D also did not occur. Callus growth was decreased, accordiong to increasing the concentration of 2.4-D. In cimbined treatments of TDZ and 2.4-D in shoot regeneration and callus proliferation, stimulated effects of cytokinin by 2.4-D did not occur entirely. Media with 0.5∼1.0mg/L TDZ ingibited the regeneration and rooting of shoots, and callus growth from callus sections. Addition of 2.4-D on medium with TDZ ingibited the regeneration and rooting of shoots, and callus growth. Rooted plantdts were acclimatized in greenhouse. The plantlets were survived more than 98% in soil of vermiculite alone or mixed perlite 1 and vermiculite 1.

• Membrane and Water Treatment
• /
• v.11 no.4
• /
• pp.283-294
• /
• 2020

This work aims at studying the feasibility of continuous removal of mixed heavy metal ions from simulated zinc plating wastewaters by coupling a microbial fuel cell and a microbial electrolysis cell in batch and continuous modes. The discharging voltage of MFC increased initially from 0.4621 ± 0.0005 V to 0.4864 ± 0.0006 V as the initial concentration of Cr⁶⁺ increased from 10 ppm to 60 ppm. Almost complete removal of Cr⁶⁺ and low removal of Cu²⁺ occurred in MFC of the MFC-MEC-coupled system after 8 hours under the batch mode; removal efficiencies (REs) of Cr⁶⁺ and Cu²⁺ were 99.76% and 30.49%. After the same reaction time, REs of nickel and zinc ions were 55.15% and 76.21% in its MEC. Cu²⁺, Ni²⁺, and Zn²⁺ removal efficiencies of 54.98%, 30.63%, 55.04%, and 75.35% were achieved in the effluent within optimum HRT of 2 hours under the continuous mode. The incomplete removal of Cu²⁺, Ni²⁺ and Zn²⁺ ions in the effluent was due to the fact that the Cr⁶⁺ was almost completely consumed at the end of MFC reaction. After HRT of 12 hours, at the different sampling locations, Cr⁶⁺ and Cu²⁺ removal efficiencies in the cathodic chamber of MFC were 89.95% and 34.69%, respectively. 94.58%, 33.95%, 56.57%, and 75.76% were achieved for Cr⁶⁺, Cu²⁺, Ni²⁺ and Zn²⁺ in the cathodic chamber of MEC. It can be concluded that those metal ions can be removed completely by repeatedly passing high concentration of Cr⁶⁺ through the cathode chamber of MFC of the MFC-MEC-coupled system.