• Science of Emotion and Sensibility
• /
• v.27 no.1
• /
• pp.59-68
• /
• 2024

This study aimed to develop a soft fabric-based elbow-bending angle sensor that can replace conventional hard-type inertial sensors and a system for estimating bending angles using it. To enhance comfort during exercise, this study treated four fabrics (Bergamo, E-band, span cushion, and polyester) by single-walled carbon nanotube dip coating to create conductive textiles. Subsequently, one fabric was selected based on performance evaluations, and an elbow flexion angle sensor was fabricated. Gauge factor, hysteresis, and sensing range were employed as performance evaluation metrics. The data obtained using the fabricated sensor showed different trends in sensor values for the changes in the angle during bending and extending movements. Because of this divergence, the two movements were separated, and this constituted the one-step process. In the two-step process, multilayer perceptron (MLP) was employed to handle the complex nonlinear relationships and achieve high data accuracy. Based on the results of this study, we anticipate effective utilization in various smart wearable and healthcare domains. Consequently, a soft- fabric bending angle sensor was developed, and using MLP, nonlinear relationships can be addressed, enabling angle estimation. Based on the results of this study, we anticipate the effective utilization of the developed system in smart wearables and healthcare.

• International conference on construction engineering and project management
• /
• 2024.07a
• /
• pp.1307-1307
• /
• 2024

Recent research on heavy equipment has been conducted for the purposes of enhanced safety, productivity improvement, and carbon neutrality at construction sites. A sensor-based approach is being explored to monitor the location and movements of heavy equipment in real time. However, it poses significant challenges in terms of time and cost as multiple sensors should be installed on numerous heavy equipment at construction sites. In addition, there is a limitation in identifying the collaboration or interference between two or more heavy equipment. In light of this, a vision-based deep learning approach is being actively conducted to effectively respond to various working conditions and dynamic environments. To enhance the performance of a vision-based activity recognition model, it is essential to secure a sufficient amount of training datasets (i.e., video datasets collected from actual construction sites). However, due to safety and security issues at construction sites, there are limitations in adequately collecting training dataset under various situations and environmental conditions. In addition, the videos feature a sequence of multiple activities of heavy equipment, making it challenging to clearly distinguish the boundaries between preceding and subsequent activities. To address these challenges, this study proposed a domain adaptation in vision-based transfer learning for automated excavator activity recognition utilizing 3D ResNet (residual deep neural network). Particularly, this study aimed to identify the optimal depth of 3D ResNet (i.e., the number of layers of the feature extractor) suitable for domain adaptation via fine-tuning process. To achieve this, this study sought to evaluate the activity recognition performance of five 3D ResNet models with 18, 34, 50, 101, and 152 layers, which used two consecutive videos with multiple activities (5 mins, 33 secs and 10 mins, 6 secs) collected from actual construction sites. First, pretrained weights from large-scale datasets (i.e., Kinetic-700 and Moment in Time (MiT)) in other domains (e.g., humans, animals, natural phenomena) were utilized. Second, five 3D ResNet models were fine-tuned using a customized dataset (14,185 clips, 60,606 secs). As an evaluation index for activity recognition model, the F1 score showed 0.881, 0.689, 0.74, 0.684, and 0.569 for the five 3D ResNet models, with the 18-layer model performing the best. This result indicated that the activity recognition models with fewer layers could be advantageous in deriving the optimal weights for the target domain (i.e., excavator activities) when fine-tuning with a limited dataset. Consequently, this study identified the optimal depth of 3D ResNet that can maintain a reliable performance in dynamic and complex construction sites, even with a limited dataset. The proposed approach is expected to contribute to the development of decision-support systems capable of systematically managing enhanced safety, productivity improvement, and carbon neutrality in the construction industry.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.1
• /
• pp.18-24
• /
• 2018

The policy-making and technological development of eco-friendly automobiles designed to increase their supply is ongoing, but the internal combustion engine still accounts for about 95% of the automobiles in use. Also, in order to meet the stricter emission regulations of internal combustion engines based on fossil fuels, the proportion of after-treatments for vehicles and (ocean going) vessels is gradually increasing. Natural gas is a clean fuel that emits few air pollutants and has been used mainly as a fuel for city buses. In the long term, we intend to develop a new NGOC/LNT+NGCO/SCR combined system that simultaneously reduces the toxic gases, $CH_4$ and NOx, emitted from CNG buses. The objective of this study is to investigate the characteristics of $de-CH_4/NOx$ according to the ceramic and metal substrates of the SCR (Selective Catalytic Reduction) catalysts mounted downstream of the combined system. The V and Cu-SCR catalysts did not affect the $CH_4$ oxidation reaction, the two NGOC/SCR catalysts each coated with two layers began to oxidize $CH_4$ at $400^{\circ}C$, and the amount of $CH_4$ emitted was reduced to about 20% of its initial value at about $550^{\circ}C$. The two NGOC/SCR catalysts each coated with two layers showed a negative (-) NOx conversion rate above $350^{\circ}C$. The ceramic-based combined system reached LOT50 at $500^{\circ}C$, which was about 20% higher in terms of the $CH_4$ conversion rate than the metal-based combined system, showing that the combined system of NGOC/LNT+Cu-SCR is a suitable combination.

• Journal of the Korean Vacuum Society
• /
• v.17 no.1
• /
• pp.34-39
• /
• 2008

We have investigated the characteristics of $Al_{0.55}Ga_{0.45}N$/GaN heterostructures with and without low-temperature (LT) AlN interlayer grown by metalorganic chemical vapor deposition. The structural and optical properties were systematically studied by Rutherford backscattering spectroscopy (RBS), X-ray diffraction (XRD), optical microscopy (OMS), scanning electron microscopy (SEM), and photoluminescence (PL). The Al content (x) of 55% and the structural properties of $Al_xGa_{1-x}N$/GaN heterostructures were investigated by using RBS and XRD, respectively. We carried out OMS and SEM experiments and obtained a decrease of the crack network in $Al_{0.55}Ga_{0.45}N$ layer with LT-AlN interlayer. A two-dimensional electron gas (2DEG)-related PL peak located at ${\sim}3.437eV$ was observed at 10 K for $Al_{0.55}Ga_{0.45}N$/GaN with LT-AlN interlayer. The 2DEG-related emission intensity gradually decreased with increasing temperature and disappeared at temperatures around 100 K. In addition, with increasing the excitation power above 3.0 mW, two 2DEG-related PL peaks were observed at ${\sim}3.411$ and ${\sim}3.437eV$. The observed lower-energy and higher-energy side 2DEG peaks were attributed to the transitions from the sub-band level and the Fermi energy level of 2DEG at the AlGaN/LT-AlN/GaN heterointerface, respectively.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.14
• /
• pp.165-172
• /
• 1973

The sterile phenomenon is frequently found in the inter-species hybrids of ginseng as in other plants. It is known that among the hybrids between Panax Ginseng (PG) and Panax Quinquefolium (PQ), and between Panax Ginseng and Paxax Japonicus (PI), PG${\times}$PI is fertile only very rarely, while PG ${\times}$ PQ is always sterile. Therefore, in order to clarify the relationship between this sterility phenomenon and the metabolism of free amino acids, the changes of free amino acids through the formation of the flower organs and seeds of two hybrids, PG ${\times}$ PQ and PG ${\times}$ PI were investigated by thin layer chromatography. The results are summarized as follows: 1. Distinct differences in the quantity and number of free amino acids were recognized between PG ${\times}$ PQ, PG ${\times}$ PI and their parent plants. From the hybrid PG ${\times}$ PQ, 19 kinds of ninhyrin sensitive substances were detected in all. They were (1) 17 amino acids: alanine, valine, leucine, phenylalanine, proline, hydroxy-proline, serine, threonine, tyrosine, aspartic acid, glutamic acid, lysine, arginine, ${\gamma}$-amino butyric acid, ${\beta}$-alanine, cysteic acid and tryptophan, and (2) two amides: asparagine and glutamine. From the hybrid PG ${\times}$ PI, in addition to the above 19 substances, methionine and one unknown substance were detected. 2. Generally, alanine, as partie acid, glutamic acid, cysteic acid and asparagine were detected in large amounts in the two hybrids as in PG, PG and PJ but it was a noticeable fact concerning these two hybrids that the largest quantity of asparagine was found at microspore satge and pollen mature stage. 3. The decrease of cysteic acid in the two hybrids at the red ripened stage was the same as in PQ and PJ but opposite to the change in PG. The detection of methionine in PG ${\times}$ PJ was worthy of notice. 4. The change of proline was conspicuously different from that in their parent plants. It was detected as a trace of color at the micros pore stage while asparagine was detected in the greatest amount at that time. It is well known that the quantity of proline is closely related to the sterility of plant. This fact was also found true in the formation of ginseng seeds. It was reported as well that asparagine accumulated when proline decreased. 5. The deficiency of proline seemed to be closely related with the sterility of hybrids and with the degradation of pollen in anther. 6. The difference in the changes of free amino acids between the selfed lines of PG, PQ and PJ, and their hybrids seemed to be caused by the transformation of gene-action system by hybridization. On these phenomena along with proline metabolim and its physiological role in seed formation further studies are required.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.24 no.3
• /
• pp.185-192
• /
• 1991

The vertical distribution and chemical characteristics of water masses were measured along two south-north transects in the polar front region of the central Korean East Sea. In February, a thermocline was present at depth between 50m and loom at the southern sites of a landward A-transect, and its depth was gradually deepened northward. At an outside B-transect, a thermocline was observed at significantly deep depth of 300m to 400m at two northern stations(Stn. 10 and 11), though the depth of the southward stations was nearly identical to that at the northward stations on a A-transect. In September, there were vertically more various water masses, i.e. the Tsushima Warm surface water(TWSW) or more than $20^{\circ}C$, the Tsushima Middle water(TMW) with a range of $12{\~}17^{\circ}C$, the North Korea Cold Water(NKCW) with $1{\~}7^{\circ}C$ temperature, the Japan Sea Proper Water(JSPW) of less than $1^{\circ}C$, and the mixed water. The North Korea Cold Water could be distinguishable from the other waters, especially from the mixed water of the Tsushima Middle Water and the Japan Sea Proper Water by the pattern of $T-O_2$ diagram. For instance, the North Korea Cold Water had higher oxygen by $1{\~}2ml/l$ than those in the mixed water, although both the two water masses ranged $1{\~}7^{\circ}C$ in water temperature. AOU value was the highest in the JSPW and the lowest in the TWSW. Also, AOU indicated a nearly linear and negative correlation with water temperature. However, AOU data for two masses, the NKCW and the TMW, in September departed remarkably from a regression line. Moreover, the ratio of $$\Delta P/\Delta AOU)$ in September was about $0.45{\mu}g-at/ml$ and higher than the value observed in the open sea. This high value could be elucidated by two factors; intrusion of the NKCW with high oxygen and molecular diffusion of dissolved oxygen from the surface into the lower layer. AOU would be a useful tracer for water masses in the polar front area of the Korean East Sea.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.139-156
• /
• 2021

The biggest reason for using a deep learning model in image classification is that it is possible to consider the relationship between each region by extracting each region's features from the overall information of the image. However, the CNN model may not be suitable for emotional image data without the image's regional features. To solve the difficulty of classifying emotion images, many researchers each year propose a CNN-based architecture suitable for emotion images. Studies on the relationship between color and human emotion were also conducted, and results were derived that different emotions are induced according to color. In studies using deep learning, there have been studies that apply color information to image subtraction classification. The case where the image's color information is additionally used than the case where the classification model is trained with only the image improves the accuracy of classifying image emotions. This study proposes two ways to increase the accuracy by incorporating the result value after the model classifies an image's emotion. Both methods improve accuracy by modifying the result value based on statistics using the color of the picture. When performing the test by finding the two-color combinations most distributed for all training data, the two-color combinations most distributed for each test data image were found. The result values were corrected according to the color combination distribution. This method weights the result value obtained after the model classifies an image's emotion by creating an expression based on the log function and the exponential function. Emotion6, classified into six emotions, and Artphoto classified into eight categories were used for the image data. Densenet169, Mnasnet, Resnet101, Resnet152, and Vgg19 architectures were used for the CNN model, and the performance evaluation was compared before and after applying the two-stage learning to the CNN model. Inspired by color psychology, which deals with the relationship between colors and emotions, when creating a model that classifies an image's sentiment, we studied how to improve accuracy by modifying the result values based on color. Sixteen colors were used: red, orange, yellow, green, blue, indigo, purple, turquoise, pink, magenta, brown, gray, silver, gold, white, and black. It has meaning. Using Scikit-learn's Clustering, the seven colors that are primarily distributed in the image are checked. Then, the RGB coordinate values of the colors from the image are compared with the RGB coordinate values of the 16 colors presented in the above data. That is, it was converted to the closest color. Suppose three or more color combinations are selected. In that case, too many color combinations occur, resulting in a problem in which the distribution is scattered, so a situation fewer influences the result value. Therefore, to solve this problem, two-color combinations were found and weighted to the model. Before training, the most distributed color combinations were found for all training data images. The distribution of color combinations for each class was stored in a Python dictionary format to be used during testing. During the test, the two-color combinations that are most distributed for each test data image are found. After that, we checked how the color combinations were distributed in the training data and corrected the result. We devised several equations to weight the result value from the model based on the extracted color as described above. The data set was randomly divided by 80:20, and the model was verified using 20% of the data as a test set. After splitting the remaining 80% of the data into five divisions to perform 5-fold cross-validation, the model was trained five times using different verification datasets. Finally, the performance was checked using the test dataset that was previously separated. Adam was used as the activation function, and the learning rate was set to 0.01. The training was performed as much as 20 epochs, and if the validation loss value did not decrease during five epochs of learning, the experiment was stopped. Early tapping was set to load the model with the best validation loss value. The classification accuracy was better when the extracted information using color properties was used together than the case using only the CNN architecture.

• Korean Journal of Medicinal Crop Science
• /
• v.8 no.3
• /
• pp.234-242
• /
• 2000

There are two cultivars, 'Jeongsun' and 'Mishima' of Bupleurum falcatum L., a medicinal plant, have been cultivated in Korea. Two cultivars were evaluated and compared in details for major growth characters and their seasonal changes. Jeongsun was higher in stem length and root weight per plant, and earlier in maturity than Mishima, while Mishima was higher in top weight per plant and the ratio of outer phloem layer(OPL) to whole root, and more in branch number. Coefficients of variations of most growth characters were higher in Mishima than in Jeongsun. In relationship between root and shoot growth characters, root weight, root diameter, etc. were positively correlated with stem thickness, top weight, branch number, etc. in both cultivars. In comparison of seasonal changes for growth characters between two cultivars, stem length and node number were almost not increased after September, but root length and branch number tended to increase until November. There were continuous increases of top and root weights in both cultivars after September. Increase rates of top and root weights were higher in Mishima and Jeongsun, respectively. The weight ratio of root part to whole plant in Jeongsun and the ratio of OPL to whole root in Mishima were higher during whole growth stages; especially Mishima showed their continuous increases after October. Jeongsun flowered earlier (August 17) than Mishima (September 16) by one month. Complete flowering date was on August 30 in Jeongsun and on September 30 in Mishima. Basic difference in plant growth between two cultivars was maturity, so shoot growth almost stopped after October in Jeongsun, and continued vegetative growth until November in Mishima with late flowering habit. These results indicate that higher root weight in Jeongsun is due to its earlier flowering character leading to relatively longer time for root growth, as compared with Mishima.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.25 no.6
• /
• pp.374-385
• /
• 2013

Sediment particle motions have been numerically simulated over a sinusoidal ripple. Turbulent boundary layer flows are generated by Large Eddy Simulation, and the sediment particle motions are simulated using Lagrangian particle tracking method. Two unsteady flow conditions are used in the experiment by employing two different wave amplitudes while keeping other conditions such as wave period same. As expected, the amount of suspended sediment particles is clearly dependent on the wave amplitude as it is increasing with increasing flow intensity. However, it is also observed that the pattern of suspension may be different as well due to the only different condition caused by wave amplitude. Specially, the time of maximum sediment suspension within the wave period is not coincident between the two cases because sediment suspension is strongly affected by the existence of turbulent eddies that are formed at different times over the ripple between the two cases as well. The role of these turbulent eddies on sediment suspension is important as it is also confirmed in previous researches. However, it is also found the time of these eddies' formation may also dependent on the wave amplitude over rippled beds. Therefore, it has been proved that various flow as well as geometric conditions under waves has to be considered in order to have better understanding on the sediment suspension process over ripples. In addition, it is found that high turbulent energy and strong upward flow velocities occur during the time of eddy formation, which also supports high suspension rate at these time steps. The results indicate that the relationship between the structure of flows and bedforms has to be carefully examined in studying sediment suspension at coastal regions.

• Korean Journal of Ecology and Environment
• /
• v.37 no.1 s.106
• /
• pp.26-35
• /
• 2004

We investigated the effect of physicochemical environmental factors on the community dynamics of phytoplanktons and bacteria at the Hoeya Dam Reservoir, a drinking water reservoir for Ulsan city. Water samples were collected and analyzed every two to four weeks at three sites along the reservoir from April to October, 2001. During the study period, the Secchi depths were between 0.4 and 3.5 m. At the surface layer of water column, temperature ranged 10.2 ~ $32.0^{\circ}C$, pH 7.3${\sim}$9.6, dissolved oxygen 5.5 ${\sim}$ 12.4 mg $L^{-1}$, $BOD_5$ 0.8 ${\sim}$ 5.0 mg $L^{-1}$, $COD_{Mn}$ 3.7 ${\sim}$ 10.0 mg $L^{-1}$, and Chl-a 8.9 ${\sim}$ 60.9 mg $m^{-3}$. At the bottom layer, temperature varied 7.2 ${\sim}$ $28.9^{\circ}C$, pH 7.1 ${\sim}$ 9.3, dissolved oxygen 0.6 ${\sim}$ 9.7 mg $L^{-1}$, $BOD_5$ 0.8 ${\sim}$ 4.5 mg $L^{-1}$, $COD_{Mn}$ 3.9 ${\sim}$ 10.0 mg $L^{-1}$, and Chl-a 4.3 ${\sim}$ 81.9 mg $m^{-3}$. The numbers of phytoplanktons were 7.4${\pm}10^2{\sim}2.6{\pm}10^5$ cells $mL^{-1}$ at surface and 2.5${\pm}10^2{\sim}2.4{\pm}10^4$ cells $mL^{-1}$ at bottom, and were positively correlated with water temperature and Chl- a concentration. Genus Stephanodiscus and genus Oscillatoria dominated on April and on May, respectively. Cyanobacterial blooms of Aphanizomenon, Microcystis, Anabaena were observed from June to early September, and thereafter Stephanodiscus and Aulacoseiral dominated again. Total microbial counts ranged 1.73${\pm}10^4{\sim}1.68{\pm}10^5$ cells $mL^{-1}$, and were positively correlated with water temperature and phytoplankton counts at surface water. Heterotrophic plate counts (HPCs) ranged 30${\sim}4.1{\pm}10^3$ CFU $mL^{-1}$, and were positively correlated with $BOD_5$ and $NO^3\;^-$-N concentration at bottom water. Unlike the total microbial counts, the numbers of fecal coliforms and fecal streptococci as well as HPCs were higher at the bottom than the surface layer and were highest at the upper a site among the three sampling sites. Since the concentrations of fecal coliforms and streptococci were still high at the bottom of site c, where intake for water treatment plant is located, it appeared that special management of water treatment processes may be needed especially after strong rainfall.