• Proceedings of the Korea Contents Association Conference
• /
• 2018.05a
• /
• pp.459-460
• /
• 2018

In this paper, we present a data augmentation method that generates synthetic plant disease images using Generative Adversarial Networks (GANs). We propose a training scheme that first uses classical data augmentation techniques to enlarge the training set and then further enlarges the data size and its diversity by applying GAN techniques for synthetic data augmentation. Our method is demonstrated on a limited dataset of 2789 images of tomato plant diseases (Gray mold, Canker, Leaf mold, Plague, Leaf miner, Whitefly etc.).

• Water for future
• /
• v.20 no.2
• /
• pp.107-116
• /
• 1987

The stochastic model has been developed for synthetic generation of hydrologic series that would be needed in the analysis, planning, design and operation of water resources system. In this study, after generating the yearly streamflows by multisite AR(1) model using the historical data in the Han River Basin, the monthly streamflows is generated by the disaggregation model. The model is verified of its applicability to domestic rivers, which is obtained through the statistical analysis and good ness of fit test using synthetic streamflows generated.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.27 no.1
• /
• pp.8-14
• /
• 2023

The existing self-supervised learning-based CutPaste generated synthetic data by cutting and attaching specific patches from normal images and then performed anomaly detection. However, this method has a problem in that there is a clear difference in the boundary of the patch. NSA for solving these problems have achieved higher anomaly detection performance by generating natural synthetic data through Poisson Blending. However, NSA has the disadvantage of having many hyperparameters that need to be adjusted for each class. In this paper, synthetic data similar to normal were generated by a simple method of making the size of the synthetic patch very small. At this time, since the patches are so locally synthesized, models that learn local features can easily overfit synthetic data. Therefore, we performed anomaly detection using gMLP, which learns global features, and even with simple synthesis methods, we were able to achieve higher performance than conventional self-supervised learning techniques.

• Water for future
• /
• v.7 no.1
• /
• pp.71-77
• /
• 1974

This paper reviews several different single site generation models for further development of a model for generating the Synthetic sequences of streamflow in the continuous streams like main streams in Korea. Initially the historical time series is looked using a time series technique, that is correlograms, to determine whether a lag one Markov model will satisfactorily represent the historical data. The single site models which were examined include an empirical model using the historical probability distribution of the random component, the linear autoregressive model(Markov model, or Thomas-Fiering model) using both logarithms of the data and Matala's log-normal transformation equations, and finally gamma distribution model.

• 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.

• Water for future
• /
• v.24 no.4
• /
• pp.59-66
• /
• 1991

Abstract The purpose of the synthetic of monthly river flows based on the short-term observed data by means of multivariate stochastic models is to provide abundunt input data to the water resources systems of which the system performance and operation policy are to be determined beforehand. In this study, multivariate Thomas-Fiering and Matalas models for synthetic generation based on stream flows in neihboring basin were employed to check if it can be applide in the modeling of monthly flows. Statistical parameters estimated by Method of Moment and Fourier Series Analysis respectively were reproduced for statistical features. For comparisons the statistical parameters of the generated monthly flow by each model were compared with those of the observed monthly flows. Results of this study suggest that the application of Matalas model for synthetic generation of monthly river flows can be adapted.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.2
• /
• pp.31-39
• /
• 2020

In this paper, we introduce a technique for generating synthetic benchmarks based on time series data. Many of the data measured on IoT devices have a time series characteristic that measures numerical changes over time. However, there is a problem that it is difficult to model the data measured over a long period as generalized time series data. To solve this problem, this paper introduces the BST-IGT model. The BST-IGT model separates the entire data into sections that can be easily time-series modeled, collects the generated data into templates, and produces new synthetic benchmarks that share or modify characteristics based on them. As a result of making a new benchmark using the proposed modeling method, we could create a benchmark with multiple aspects by mixing the composite benchmark with the statistical features of the existing data and other benchmarks.

• Korean Journal of Clinical Pharmacy
• /
• v.11 no.2
• /
• pp.68-77
• /
• 2001

Herbal medicines have been used since ancient times as medicines to treat and relieve the symptoms of many different human diseases. However, so far, relatively few herbal medicines have been evaluated scientifically to prove their safety, potential benefits and effectiveness. This study was conducted to provide the groundwork for improving the current registration regulations for herbal medicines in distinction from synthetic medicines. The study was performed based on the literature research and individual interviews with 5 experts who had extensive experience in registration of herbal medicines. When compared with synthetic drugs, herbal medicines exhibit some marked differences, namely the active principles are frequently unknown, standardization, stability and quality control are not easy, they are usually mixtures of complex compounds. Second, the current regulations for herbal medicines are reviewed by comparison of foreign regulation systems like the one in China. The regulation requirements of herbal medicine in China are in distinction from synthetic drugs. The authors conclude that new registration requirements for the herbal medicines should be changed as follows; the toxicity and efficacy data should be submitted as mixed herbal preparation and the documents and other research on the reproduction and generation toxicity need to be shown for the proof of reproduction and generation toxicity, if needed.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.37 no.2
• /
• pp.31-40
• /
• 1995

This study was conducted to get best fitting frequency distribution for the annual run- off and to simulate long series of annual flows by single-season first order Markov Model with comparison of statistical parameters which were derived from observed and synthetic flows at four watersheds in Seom Jin and Yeong San river systems. The results summarized through this study are as follows. 1. Hydrologic persistence of observed flows was acknowledged by the correlogram analysis. 2. A normal distribution of the annual runoff for the applied watersheds was confirmed as the best one among others by Kolmogorov-Smirnov test. 3. Statistical parameters were calculated from synthetic flows simulated by normal dis- tribution. In was confirmed that mean and standard deviation of simulated flows are much closer to those of observed data than except coefficient of skewness. 4. Hydrologic persistence between observed flows and synthetic flows simulated was also confirmed by the correlogram analysis. 5. It is to be desired that generation technique of synthetic flow in this study would be compared with other simulation techniques for the objective time series.

• Proceedings of The Korean Society of Agricultural and Forest Meteorology Conference
• /
• 2016.09a
• /
• pp.25-28
• /
• 2016

Recently, because of the weather forecasts through the low-resolution data has been limited, the demand of the high-resolution data is sharply increasing. Therefore, in this study, we restore the ultra-high resolution synthetic precipitation and temperature data for 2000-2014 due to small-scale topographic effect using the QPM (Quantitative Precipitation Model)/QTM (Quantitative Temperature Model). First, we reproduce the detailed precipitation and temperature data with 1km resolution using the distribution of Automatic Weather System (AWS) data and Automatic Synoptic Observation System (ASOS) data, which is about 10km resolution with irregular grid over South Korea. Also, we recover the precipitation and temperature data with 1km resolution using the MERRA reanalysis data over North Korea, because there are insufficient observation data. The precipitation and temperature from restored current climate reflect more detailed topographic effect than irregular AWS/ASOS data and MERRA reanalysis data over the Korean peninsula. Based on this analysis, more detailed prospect of regional climate is investigated.