• Journal of the Korea Society of Computer and Information
• /
• v.29 no.1
• /
• pp.31-40
• /
• 2024

In the realm of domestic online fashion platform industry the manual registration of product information by individual business owners leads to inconvenience and reliability issues, especially when dealing with simultaneous registrations of numerous product groups. Moreover, bias is significantly heightened due to the low quality of product images and an imbalance in data quantity. Therefore, this study proposes a ResNet50 model aimed at minimizing data bias through oversampling techniques and conducting multiple classifications for 13 fashion categories. Transfer learning is employed to optimize resource utilization and reduce prolonged learning times. The results indicate improved discrimination of up to 33.4% for data augmentation in classes with insufficient data compared to the basic convolution neural network (CNN) model. The reliability of all outcomes is underscored by precision and affirmed by the recall curve. This study is suggested to advance the development of the domestic online fashion platform industry to a higher echelon.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.6
• /
• pp.566-571
• /
• 2009

This paper proposes noise-robust fast speaker adaptation method based on the eigenvoice framework in various noisy environments. The proposed method is focused on de-noising and environment clustering. Since the de-noised adaptation DB still has residual noise in itself, environment clustering divides the noisy adaptation data into similar environments by a clustering method using the cepstral mean of non-speech segments as a feature vector. Then each adaptation data in the same cluster is used to build an environment-clustered speaker adapted (SA) model. After selecting multiple environmentally clustered SA models which are similar to test environment, the speaker adaptation based on an appropriate linear combination of clustered SA models is conducted. According to our experiments, we observe that the proposed method provides error rate reduction of $40{\sim}59%$ over baseline with speaker independent model.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.48 no.2
• /
• pp.312-327
• /
• 2024

In this study, a rehabilitation 3D printed wearable device was developed by combining an assembly-type robot hand and an integral-type robot hand through fused deposition 3D printing manufacturing with various hardness TPU (Thermoplastic Polyurethane) filaments. The hardware configuration of the robot hand includes a controller designed with four motors, one small servo motor, and a circuit board. In the case of the assembly-type robot hand model, a 3D printed robot hand was assembled using samples printed with TPU of hardness 87A and 95A. It was observed that TPU with a hardness of 95A was suitable for use due to shape stability. For the integrated-type robot hand model, the external sample using TPU of hardness 95A could be modified through a cutting method, and the hardware configuration is the same as the assembly-type. The system structure of the 3D printed robot hand was improved from an individual control method to a simultaneous transmission method.Furthermore, the system architecture of an integrated 3D printed robotic hand rehabilitation device and the application of the rehabilitation device were developed.

• Journal of Korean Dental Science
• /
• v.17 no.2
• /
• pp.53-63
• /
• 2024

Purpose: To investigate the effects of simultaneous soft and hard tissue augmentation and the addition of polydeoxyribonucleotide (PDRN) on regenerative outcomes. Materials and Methods: In five mongrel dogs, chronic ridge defects were established in both mandibles. Six implants were placed in the mandible, producing buccal dehiscence defects. The implants were randomly allocated to one of the following groups: 1) control: no treatment; 2) GBR: guided bone regeneration (GBR) only; 3) GBR/PDRN: GBR+PDRN application to bone substitute particles; 4) GBR/CTG: GBR+connective tissue grafting (CTG); 5) GBR/VCMX: GBR+soft tissue augmentation using volume stable collagen matrix (VCMX); and 6) group GBR/VCMX/PDRN: GBR+VCMX soaked with PDRN. The healing abutments were connected to the implants to provide additional room for tissue regeneration. Submerged healing was achieved. The animals were euthanized after four months. Histological and histomorphometric analyses were then performed. Results: Healing abutments were gradually exposed during the healing period. Histologically, minimal new bone formation was observed in the dehiscence defects. No specific differences were found between the groups regarding collagen fiber orientation and density in the augmented area. No traces of CTG or VCMX were detected. Histomorphometrically, the mean tissue thickness was greater in the control group than in the other groups above the implant shoulder (IS). Below the IS level, the CTG and PDRN groups exhibited more favorable tissue thickness than the other groups. Conclusion: Failure of submerged healing after tissue augmentation deteriorated the tissue contour. PDRN appears to have a positive effect on soft tissues.

• Korean Journal of Remote Sensing
• /
• v.35 no.6_3
• /
• pp.1209-1219
• /
• 2019

For consistent vegetation monitoring, it is necessary to generate time-series vegetation index datasets at fine temporal and spatial scales by fusing the complementary characteristics between temporal and spatial scales of multiple satellite data. In this study, we quantitatively and qualitatively analyzed the prediction accuracy of time-series change information extracted from spatio-temporal fusion models of multiple satellite data for vegetation monitoring. As for the spatio-temporal fusion models, we applied two models that have been widely employed to vegetation monitoring, including a Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and an Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM). To quantitatively evaluate the prediction accuracy, we first generated simulated data sets from MODIS data with fine temporal scales and then used them as inputs for the spatio-temporal fusion models. We observed from the comparative experiment that ESTARFM showed better prediction performance than STARFM, but the prediction performance for the two models became degraded as the difference between the prediction date and the simultaneous acquisition date of the input data increased. This result indicates that multiple data acquired close to the prediction date should be used to improve the prediction accuracy. When considering the limited availability of optical images, it is necessary to develop an advanced spatio-temporal model that can reflect the suggestions of this study for vegetation monitoring.

• The KIPS Transactions:PartB
• /
• v.12B no.3 s.99
• /
• pp.349-356
• /
• 2005

Artificial Organism-used application areas are expanding at a break-neck speed with a view to getting things done in a dynamic and Informal environment. A use of general programming or traditional hi methods as the representation of Artificial Organism behavior knowledge in these areas can cause problems related to frequent modifications and bad response in an unpredictable situation. Strategies aimed at solving these problems in a machine-learning fashion includes Genetic Programming and Evolving Neural Networks. But the learning method of Artificial-Organism is not good yet, and can't represent life in the environment. With this in mind, this research is designed to come up with a new behavior evolution model. The model represents behavior knowledge with Classification Rules and Enhanced Backpropation Neural Networks and discriminate the denomination. To evaluate the model, the researcher applied it to problems with the competition of Artificial-Organism in the Simulator and compared with other system. The survey shows that the model prevails in terms of the speed and Qualify of learning. The model is characterized by the simultaneous learning of classification rules and neural networks represented on chromosomes with the help of Genetic Algorithm and the consolidation of learning ability caused by the hybrid processing of the classification rules and Enhanced Backpropagation Neural Network.

• Journal of the Korea Convergence Society
• /
• v.13 no.2
• /
• pp.67-75
• /
• 2022

The purpose of this study is to present an appropriate management plan as a supplement to the scientific evidence of the currently operated distancing system for preventing COVID-19. The currently being used mathematical models are expressed as simultaneous ordinary differential equations, there is a problem in that it is difficult to use them for the management of entry and exit of small business owners. In order to supplement this point, in this paper, a method for quantitatively expressing the risk of infection by people who gather is presented in consideration of the allowable risk given to the gathering space, the basic infection reproduction index, and the risk reduction rate due to vaccination. A simple quantitative model was developed that manages the probability of infection in a probabilistic level according to a set of visitors by considering both the degree of infection risk according to the vaccination status (non-vaccinated, primary inoculation, and complete vaccination) and the epidemic status of the virus. In a given example using the model, the risk was reduced to 55% when 20% of non-vaccinated people were converted to full vaccination. It was suggested that management in terms of quarantine can obtain a greater effect than medical treatment. Based on this, a generalized model that can be applied to various situations in consideration of the type of vaccination and the degree of occurrence of confirmed cases was also presented. This model can be used to manage the total risk of people gathered at a certain space in a real time, by calculating individual risk according to the type of vaccine, the degree of inoculation, and the lapse of time after inoculation.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.29 no.5B
• /
• pp.397-408
• /
• 2009

In this study, a stochastic precipitation generation framework for simultaneous simulation of daily precipitation at multiple sites is presented. The precipitation occurrence at individual sites is generated using hybrid-order Markov chain model which allows higher-order dependence for dry sequences. The precipitation amounts are reproduced using Anscombe residuals and gamma distributions. Multisite spatial correlations in the precipitation occurrence and amount series are represented with spatially correlated random numbers. The proposed model is applied for a network of 17 locations in the middle of Korean peninsular. Evaluation statistics are reported by generating 50 realizations of the precipitation of length equal to the observed record. The analysis of results show that the model reproduces wet day number, wet and dry day spell, and mean and standard deviation of wet day amount fairly well. However, mean values of 50 realizations of generated precipitation series yield around 23% Root Mean Square Errors (RMSE) of the average value of observed maximum numbers of consecutive wet and dry days and 17% RMSE of the average value of observed annual maximum precipitations for return periods of 100 and 200 years. The provided model also reproduces spatial correlations in observed precipitation occurrence and amount series accurately.

• Journal of the Korea Society for Simulation
• /
• v.32 no.4
• /
• pp.41-58
• /
• 2023

As naval warfare changes to composite warfare that includes simultaneous engagements against surface, underwater, and air enemies, performance and tactical analysis are required to respond to naval warfare. In particular, for practical analysis of composite warfare, it is necessary to study engagement simulations that can appropriately utilize the limited performance resources of the detection system. This paper proposes a DEVS (Discrete Event Systems Specifications)-based simulation model for composite warfare analysis. The proposed model contains generalized models of combat platforms and armed objects to simulate various complex warfare situations. In addition, we propose a detection performance allocation algorithm that can be applied to a detection system model, considering the characteristics of composite warfare in which missions must be performed using limited detection resources. We experimented with the effectiveness of composite warfare according to the strength of the detection system's resource allocation, the enemy force's size, and the friendly force's departure location. The simulation results showed the effect of the resource allocation function on engagement time and success. Our model will be used as an engineering basis for analyzing the tactics of warships in various complex warfare situations in the future.

• KDI Journal of Economic Policy
• /
• v.28 no.2
• /
• pp.131-177
• /
• 2006

This paper examines a simultaneous determination of corporate control mechanisms, and its effects on firm performance. The corporate control mechanisms considered include the following; insider shareholding, institutional shareholding, the board of directors, dividend policy, and capital structure. This paper applies a simultaneous equation methodology and investigates the interdependence among the corporate control mechanisms. In the first part, the paper finds that firm-level variations of control mechanisms are large across time although average variations are relatively small. These variations are related to one another, which is confirmed by Granger causality test based on dynamic panel autoregression model. More specifically insider shareholding, institutional shareholding and outside director ratio cause each other. With regard to interdependence among the control mechanisms, 2SLS(two stage least squares) regression results show that insider shareholding and institutional shareholding are substitutes while institutional shareholding acts as complements to the ratio of outside members in the board of directors. Then in the second part, the paper examines the relationship between firm performance and corporate governance. Firm performance, measured by Tobin's Q, has a positive association with leverage ratio while that has a negative relation to outside director ratio. This suggests that there may be a room for reforming corporate governance in Korea. Specifically it is necessary to enhance the independence of the outside directors.