• Journal of Intelligence and Information Systems
• /
• v.28 no.4
• /
• pp.179-190
• /
• 2022

Data imbalance refers to a phenomenon in which the number of data in one category is too large or too small compared to another category. Due to this, it has been raised as a major factor that deteriorates performance in machine learning that utilizes classification algorithms. In order to solve the data imbalance problem, various ovrsampling methods for amplifying prime number distribution data have been proposed. Among them, SMOTE is the most representative method. In order to maximize the amplification effect of minority distribution data, various methods have emerged that remove noise included in data (SMOTE-IPF) or enhance only border lines (Borderline SMOTE). This paper proposes a method to ultimately improve classification performance by improving the processing method for anomaly data in the traditional SMOTE method that amplifies minority classification data. The proposed method consistently presented relatively high classification performance compared to the existing methods through experiments.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.13 no.1
• /
• pp.379-395
• /
• 2021

For the complicated structural details in ships and offshore structures, the traditional hotspot stress approaches are known to be sensitive to the element variables of element topologies, sizes, and integration schemes. This motivated to develop a new approach for predicting reasonable hotspot stresses, which is less sensitive to the element variables and easy to be implemented the real marine structures. The three-point bending tests were conducted for the longitudinal attachments with the round and rectangular weld toes. The tests were reproduced in the numerical simulations using the solid and shell element models, and the simulation technique was validated by comparing the experimental stresses with the simulated ones. This paper considered three hotspot stress approaches: the ESM method based on surface stress extrapolation, the Dong's method based on nodal forces along a weld toe, and the proposed method based on nodal forces perpendicular to an imaginary vertical plane at a weld toe. In order to study the element sensitivities of each method, 16 solid element models and 8 shell element models were generated under the bending and tension loads, respectively. The element sensitivity was analyzed in terms of Stress Concentration Factors (SCFs) in viewpoints of two statistical quantities of mean and bias with respect to the reference SCFs. The average SCFs predicted by the proposed method were remarkably in good agreement with the reference SCFs based on the experiments and the ship rules. Negligibly small Coefficients of Variation (CVs) of the SCFs, which is measure of statistical bias, were drawn by the proposed method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.12
• /
• pp.4420-4438
• /
• 2021

Image dehazing is an ill-posed problem which is far from being solved. Traditional image dehazing methods often yield mediocre effects and possess substandard processing speed, while modern deep learning methods perform best only in certain datasets. The haze removal effect when processed by said methods is unsatisfactory, meaning the generalization performance fails to meet the requirements. Concurrently, due to the limited processing speed, most dehazing algorithms cannot be employed in the industry. To alleviate said problems, a lightweight fast dehazing network based on a multiple scale-patch framework (MSP) is proposed in the present paper. Firstly, the multi-scale structure is employed as the backbone network and the multi-patch structure as the supplementary network. Dehazing through a single network causes problems, such as loss of object details and color in some image areas, the multi-patch structure was employed for MSP as an information supplement. In the algorithm image processing module, the image is segmented up and down for processed separately. Secondly, MSP generates a clear dehazing effect and significant robustness when targeting real-world homogeneous and nonhomogeneous hazy maps and different datasets. Compared with existing dehazing methods, MSP demonstrated a fast inference speed and the feasibility of real-time processing. The overall size and model parameters of the entire dehazing model are 20.75M and 6.8M, and the processing time for the single image is 0.026s. Experiments on NTIRE 2018 and NTIRE 2020 demonstrate that MSP can achieve superior performance among the state-of-the-art methods, such as PSNR, SSIM, LPIPS, and individual subjective evaluation.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.2
• /
• pp.81-86
• /
• 2022

Given that time series are used in various fields, such as finance, IoT, and manufacturing, data analytical methods for accurate time-series forecasting can serve to increase operational efficiency. Among time-series analysis methods, multi-horizon forecasting provides a better understanding of data because it can extract meaningful statistics and other characteristics of the entire time-series. Furthermore, time-series data with exogenous information can be accurately predicted by using multi-horizon forecasting methods. However, traditional deep learning-based models for time-series do not account for the heterogeneity of inputs. We proposed an improved time-series predicting method, called the temporal fusion transformer method, which combines multi-horizon forecasting with interpretable insights into temporal dynamics. Various real-world data such as stock prices, fine dust concentrates and electricity consumption were considered in experiments. Experimental results showed that our temporal fusion transformer method has better time-series forecasting performance than existing models.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.27 no.7
• /
• pp.1129-1136
• /
• 2021

A pan-tilt-zoom (PTZ) camera-based monitoring system for efficient monitoring in the engine room of a vessel was designed. A number of places exist where traditional analog instruments are still used in vessel engine rooms, and blind spots closely related to safety exist, for which flooding or fire is a concern. A camera-based monitoring system that guarantees a wide range at a relatively fast cycle for these monitoring points can be an effective alternative to enhance the safety of a vessel. Therefore, a multiview monitoring system is proposed in which the functions of the existing PTZ camera are further strengthened using a software. The monitoring system comprises four modules: camera control, location registration, traversal control, and multiview image reconstruction. The effectiveness of the method was evaluated through a series of experiments in an engine room environment.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.4
• /
• pp.577-586
• /
• 2022

Detection and classification of steel surface defects are critical for product quality control in the steel industry. However, due to its low accuracy and slow speed, the traditional approach cannot be effectively used in a production line. The current, widely used algorithm (based on deep learning) has an accuracy problem, and there are still rooms for development. This paper proposes a method of steel surface defect detection combining EfficientNetV2 for image classification and YOLOv5 as an object detector. Shorter training time and high accuracy are advantages of this model. Firstly, the image input into EfficientNetV2 model classifies defect classes and predicts probability of having defects. If the probability of having a defect is less than 0.25, the algorithm directly recognizes that the sample has no defects. Otherwise, the samples are further input into YOLOv5 to accomplish the defect detection process on the metal surface. Experiments show that proposed model has good performance on the NEU dataset with an accuracy of 98.3%. Simultaneously, the average training speed is shorter than other models.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.3
• /
• pp.93-98
• /
• 2022

Recently, fast storage media such as Optane have emerged, and memory system configurations designed for disk storage should be reconsidered. In this paper, we analyze the effect of the page size on the memory system performances when fast storage is adopted. Based on this, we design a page size model that can guide an appropriate page size for given workloads in virtualized environments. Configuring different page sizes for various workloads is not an easy matter in traditional systems, but due to the widespread adoption of cloud systems, page sizing performed in our model is feasible for virtual machines, which are generated for executing specific workloads. Simulation experiments under various virtual machine scenarios show that the proposed model improves the memory access time significantly by configuring page sizes for given workloads.

• Smart Structures and Systems
• /
• v.31 no.2
• /
• pp.113-130
• /
• 2023

Hybrid simulation (HS) has attracted community attention in recent years as an efficient and effective experimental technique for structural performance evaluation in size-limited laboratories. Traditional hybrid simulations usually take deterministic properties for their numerical substructures therefore could not account for inherent uncertainties within the engineering structures to provide probabilistic performance assessment. Reliable structural performance evaluation, therefore, calls for stochastic hybrid simulation (SHS) to explicitly account for substructure uncertainties. The experimental design of SHS is explored in this study to account for uncertainties within analytical substructures. Both computational simulation and laboratory experiments are conducted to evaluate the pseudo-random Sobol sequence for the experimental design of SHS. Meta-modeling through polynomial chaos expansion (PCE) is established from a computational simulation of a nonlinear single-degree-of-freedom (SDOF) structure to evaluate the influence of nonlinear behavior and ground motions uncertainties. A series of hybrid simulations are further conducted in the laboratory to validate the findings from computational analysis. It is shown that the Sobol sequence provides a good starting point for the experimental design of stochastic hybrid simulation. However, nonlinear structural behavior involving stiffness and strength degradation could significantly increase the number of hybrid simulations to acquire accurate statistical estimation for the structural response of interests. Compared with the statistical moments calculated directly from hybrid simulations in the laboratory, the meta-model through PCE gives more accurate estimation, therefore, providing a more effective way for uncertainty quantification.

• The Korean Journal of Physiology and Pharmacology
• /
• v.27 no.1
• /
• pp.1-8
• /
• 2023

Hypothyroidism alone can lead to myocardial fibrosis and result in heart failure, but traditional hormone replacement therapy does not improve the fibrotic situation. Hydrogen sulfide (H₂S), a new gas signaling molecule, possesses anti-inflammatory, antioxidant, and anti-fibrotic capabilities. Whether H₂S could improve hypothyroidism-induced myocardial fibrosis are not yet studied. In our study, H₂S could decrease collagen deposition in the myocardial tissue of rats caused by hypothyroidism. Furthermore, in hypothyroidism-induced rats, we found that H₂S could enhance cystathionine-gamma-lyase (CSE), not cystathionine β-synthase (CBS), protein expressions. Finally, we noticed that H₂S could elevate autophagy levels and inhibit the transforming growth factor-β1 (TGF-β1) signal transduction pathway. In conclusion, our experiments not only suggest that H₂S could alleviate hypothyroidism-induced myocardial fibrosis by activating autophagy and suppressing TGF-β1/SMAD family member 2 (Smad 2) signal transduction pathway, but also show that it can be used as a complementary treatment to conventional hormone therapy.

• Journal of Korean Medicine for Obesity Research
• /
• v.22 no.2
• /
• pp.147-157
• /
• 2022

Objectives: The purpose of this study is to research domestic and foreign trend of pharmacopuncture treatment on obesity during recent 10 years. Methods: 5 Databases (Korean Studies Information Service System, Research Information Sharing Service, Oriental Medicine Advanced Searching Integrated System, Scopus, PubMed) were searched with keywords of ('pharmacopuncture', 'herbal acupuncture', 'aquapuncture', 'obesity') from 2012 to 2022. Results: 25 Articles were selected and analyzed. 15 articles (60%) were animal experimentations, 8 articles (32%) were case reports, 1 article (4%) was cell experimentation, and 1 article (4%) was clinical trial. In this study, 25 articles were analyzed by subject, acupoints, injections, metrics and results. Pharmacopuncture treatment for obesity is continuously being studied, and the anti-inflammatory effect as well as the effect of reducing obesity factors has been revealed. Conclusions: This study suggests the efficacy and future development of pharmacopuncture for obesity. The studies of the past decade have been concentrated on animal experiments, so many clinical trials and various studies on new complex pharmacopuncture for obesity are expected.