• Journal of Korea Multimedia Society
• /
• v.23 no.2
• /
• pp.186-203
• /
• 2020

Image-based Virtual Try-On Systems are among the most potential solution for virtual fitting which tries on a target clothes into a model person image and thus have attracted considerable research efforts. In many cases, current solutions for those fails in achieving naturally looking virtual fitted image where a target clothes is transferred into the body area of a model person of any shape and pose while keeping clothes context like texture, text, logo without distortion and artifacts. In this paper, we propose a new improved image-based virtual try-on network system based on keypoints, which we name as KP-VTON. The proposed KP-VTON first detects keypoints in the target clothes and reliably predicts keypoints in the clothes of a model person image by utilizing a dense human pose estimation. Then, through TPS transformation calculated by utilizing the keypoints as control points, the warped target clothes image, which is matched into the body area for wearing the target clothes, is obtained. Finally, a new try-on module adopting Attention U-Net is applied to handle more detailed synthesis of virtual fitted image. Extensive experiments on a well-known dataset show that the proposed KP-VTON performs better the state-of-the-art virtual try-on systems.

• THE INTERNATIONAL COMMERCE & LAW REVIEW
• /
• v.17
• /
• pp.93-125
• /
• 2002

This paper aims to find a reasonable solvency margin system in non-life insurance industry and also to evaluate the appropriateness of the current solvency margin regulation system in Korea. The current solvency margin system in Korea, based on EU's solvency margin model, was introduced during the 1997 financial crisis. The solvency requirement is not based on non-life insurer's risk, but simply on written premiums. The current solvency margin for general insurance, such as fire, marine, and automobile insurance, is determined by the greater between a premium-based amount and a claim-based amount, where the premium-based solvency margin is calculated by multiplying the net written premium for the preceding year by the premium based solvency margin ratio. Also, the amount of solvency margin for long term insurance is set at 4% of the policy reserve of the long term insurance. Still, there exist many differences between the current solvency margin regulation system in Korea and EU's model. This paper focuses on the rationality of the solvency margin regulation system, and compares the current system in Korea with EU's model and the RBC(Risk Based Capital) system in U.S. and Japan. Finally, this paper suggests a more specific and reasonable solvency margin system to be developed in Korea.

• The Journal of Asian Finance, Economics and Business
• /
• v.5 no.1
• /
• pp.5-9
• /
• 2018

Many developing countries have attempted to depreciate their currencies in order to make their products cheaper, stimulate exports, shift aggregate demand to the right, and increase aggregate output. However, currency depreciation tends to increase import prices, raise domestic inflation, reduce capital inflows, and shift aggregate supply to the left. The net impact is unclear. The paper incorporates the monetary policy function in the model, which is determined by the inflation gap, the output gap, the real effective exchange rate, and the world real interest rate. Applying an extended IS-MP-AS model (Romer, 2000), the paper finds that real depreciation raised real GDP during 1997.Q1-2005.Q3 whereas real appreciation increased real GDP during 2005.Q4-2017.Q2. In addition, a higher government debt-to-GDP ratio, a lower U.S. real federal funds rate, a higher real stock price, a lower real oil price or a lower expected inflation rate would help increase real GDP. Hence, real depreciation or real appreciation may increase or reduce aggregate output, depending upon the level of economic development. Although expansionary fiscal policy is effective in stimulating the economy, caution needs to be exercised as there may be a debt threshold beyond which a further increase in the debt-to-GDO ratio would hurt economic growth.

• Solar Energy
• /
• v.20 no.3
• /
• pp.39-50
• /
• 2000

Radiant barrier systems(RES) constructed with low emissivity materials bounded by an open air space can be used to reduce the net radiation transfer between two surfaces. To analyze the heat transfer characteristics of the radiant barrier systems which consist of a single-glass and radiation barriers, a simple theoretical model based on energy balances was suggested. And the model was validated by means of the experimental results. Using a guarded hot box, the temperatures of layers in selected RES and energy use for each cases were measured. The results show that the model well explained the heat transfer characteristics of those RES. Also, the heat transfer coefficient correlations considering natural and forced convection heat transfer ware suggested. It is found that the heat transfer efficiency of a RBS with aluminium surface improved up to 66.6% over that of a single glazing system.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.10a
• /
• pp.637-640
• /
• 2018

The paper proposes a new deep network-based model that rates periorbital wrinkles in order to alleviate the shortcomings of the evaluation by human experts as well as to facilitate the automation. Periorbital wrinkles still need to be classified by human experts. Furthermore, the classification results from experts are different from each other in many cases due to the inter-interpreter variability and the absence of quantification criteria. Unlike existing classification methods which classify original images, the proposed model consists of a cascade of two deep networks: U-Net for the enhancement of wrinkles on an input image and VGG16 for final classification based on the wrinkle information. Experiments of the proposed model are made with a data set that consists of 433 images rated by experts, showing the promising performance.

• Nuclear Engineering and Technology
• /
• v.52 no.9
• /
• pp.1955-1962
• /
• 2020

In a sudden shutdown of primary pump or coolant loss accident in a marine nuclear power plant, the primary flow decreases rapidly in a transition process from forced circulation (FC) to natural circulation (NC), and the lower flow enters the steam generator (SG) causing reverse flow in the U-tube. This can significantly compromise the safety of nuclear power plants. Based on the marine natural circulation steam generator (NCSG), an experimental loop is constructed to study the characteristics of reverse flow under middle-temperature and middle-pressure conditions. The transition from FC to NC is simulated experimentally, and the characteristics of SG reverse flow are studied. On this basis, the experimental loop is numerically modeled using RELAP5/MOD3.3 code for system analysis, and the accuracy of the model is verified according to the experimental data. The influence of the flow variation rate on the reverse flow phenomenon and flow distribution is investigated. The experimental and numerical results show that in comparison with the case of adjusting the mass flow discontinuously, the number of reverse flow tubes increases significantly during the transition from FC to NC, and the reverse flow has a more severe impact on the operating characteristics of the SG. With the increase of flow variation rate, the reverse flow is less likely to occur. The mass flow in the reverse flow U-tubes increases at first and then decreases. When the system is approximately stable, the reverse flow is slightly lower than obverse flow in the same U-tube, while the flow in the obverse flow U-tube increases.

• Geophysics and Geophysical Exploration
• /
• v.24 no.3
• /
• pp.89-97
• /
• 2021

Fault detection in seismic data is well suited to the application of machine learning algorithms. Accordingly, various machine learning techniques are being developed. In recent studies, machine learning models, which utilize synthetic data, are the particular focus when training with deep learning. The use of synthetic training data has many advantages; Securing massive data for training becomes easy and generating exact fault labels is possible with the help of synthetic training data. To interpret real data with the model trained by synthetic data, the synthetic data used for training should be geologically realistic. In this study, we introduce a method to generate realistic synthetic seismic data. Initially, reflectivity models are generated to include realistic fault structures, and then, a one-way wave equation is applied to efficiently generate seismic stack sections. Next, a migration algorithm is used to remove diffraction artifacts and random noise is added to mimic actual field data. A convolutional neural network model based on the U-Net structure is used to verify the generated synthetic data set. From the results of the experiment, we confirm that realistic synthetic data effectively creates a deep learning model that can be applied to field data.

• Geotechnical Engineering
• /
• v.14 no.3
• /
• pp.63-72
• /
• 1998

Laboratory model test results for ultimate uplift capacity of horizontal circular anchors embedded in soft clay overlain by dense sand are presented. The effect of the critical embedment ratio on the thickness of the clay layer was evalyated. An approximate preocedure for estimating the net ultimate capacity is presented.

• International Journal of Human Ecology
• /
• v.9 no.1
• /
• pp.81-95
• /
• 2008

This paper examined the impact of psychological characteristics of consumers on household debt decisions. With the use of the Heckit models (the traditional approach to the selection problem) this study undertook an empirical study of the influence of a wide range of factors on financial decisions. This study used U.S. household-level data that offers detailed information on household debt, expectations about future income, expectations about future economic conditions, the amount of financial risk the respondent was willing to take, and the amount of time allotted for planning family savings and spending. This study showed that respondents with both substantial financial risk tolerance and positive expectations about future income were likely to have larger household debt showing that researchers and policy-makers need to consider consumer sentiment and preference measures in modeling behavior in credit markets. Additional results showed that household debt is significantly related to two key economic variables: income and net worth.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2005.11a
• /
• pp.346-355
• /
• 2005

In this paper, we propose a meta-search engine which provides distributed product information through a unified access to multiple e-commerce. The meta-search engine proposed in this paper performs the following functions: (I) The user is able to create a category-based user query, (2) by using the WordNet, the query is semantical refined fined for increasing search accuracy, and (3) the meta-search engine recommends an e-commerce site which has the closest product information to the user's search intention, by matching the user query with the product catalogs in the e-commerce sites linked to the meta-search engine. An experiment shows that the performance of our model is better than that of general keyword-based search.