• 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.).

• Journal of Information Technology and Architecture
• /
• v.9 no.3
• /
• pp.253-263
• /
• 2012

Presented in the paper is a methodology for the distributed simulation of underwater warfare based on standard synthetic environment. In the case of underwater warfare simulation, it is very important to reflect environmental data, such as salinity and temperature. For the reusability and interoperability of environmental data, this paper adopts Synthetic Environmental Data Representation and Interchange Specification (SEDRIS(ISO standard for environmental data)). Although SEDRIS provides various merits as an international standard, applying of SEDRIS has been hindered by its broadness and heaviness. To relieve the difficulties, this paper proposes an efficient procedure to utilize SEDRIS technology for the atmosphere and underwater environment. This paper identifies SEDRIS structure for the atmosphere/underwater structured dimensional grid-based and implements the proposed procedure on the High Level Architecture (HLA) / Run-Time Infrastructure (RTI) to explain the generation of the battlefield in a distributed simulation system.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.1
• /
• pp.130-138
• /
• 2003

The classical image reconstruction for stripmap-mode Synthetic Aperture Radar is the Range-Doppler algorithm. When the spotlight-mode SAR system was envisioned, Range-Doppler algorithm turned out to fail rapidly in this SAR imaging modality. Thus, what is referred to as Polar format algorithm, which is based on the Plane wave approximation, was introduced for imaging from spotlight-mode SAR raw- data. In this paper, we have studied for the raw data processing schemes in the spotlight-mode Synthetic Aperture Radar. We apply the Wavefront Reconstruction scheme that does not utilize the approximation in spotlight-mode SAR imaging modelity, and compare the performance of target imaging with the Polar format inversion scheme.

• Proceedings of the IEEK Conference
• /
• 2000.11a
• /
• pp.501-504
• /
• 2000

The classical image reconstruction for stripmap SAR is the range-Doppler imaging. However, when the spotlight SAR system was envisioned, range-Bowler imaging fumed out to fail rapidly in this SAR imaging modality. What is referred to as polar format processing, which is based on the plane wave approximation, was introduced for imaging from spotlight SAR data. This paper has been studied for the raw data processing schemes in the spotlight-mode synthetic aperture radar. we apply the wavefront reconstruction scheme that does not utilize the approximation in spotlight-mode SAR imaging modelity, and compare the performance of target imaging with the polar format inversion scheme.

• Journal of The Korean Astronomical Society
• /
• v.37 no.5
• /
• pp.593-596
• /
• 2004

We introduce a method of identifying evidence of shocks in the X-ray emitting gas in clusters of galaxies. Using information from synthetic observations of simulated clusters, we do a blind search of the synthetic image plane. The locations of likely shocks found using this method closely match those of shocks identified in the simulation hydrodynamic data. Though this method assumes nothing about the geometry of the shocks, the general distribution of shocks as a function of Mach number in the cluster hydrodynamic data can be extracted via this method. Characterization of the cluster shock distribution is critical to understanding production of cosmic rays in clusters and the use of shocks as dynamical tracers.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.35 no.3
• /
• pp.100-116
• /
• 1993

It is experienced fact that unreasonable design criterion and unsitable operation management for the agricultural structures including reservoirs based on short terms data of monthly flows have been brought about not only loss of lives, but also enormous property damage. For the solution of this point at issue, this study was conducted to simulate long series of synthetic monthly flows by multi-season first order Markov model with selection of best fitting frequency distribution and to make a comparison of statistical parameters between observed and synthetic flows of six watersheds in Yeong San and Seom Jin river systems. The results obtained through this study can be summarized as follows. 1.Both Gamma and two parameter lognormal distribution were found to be suitable ones for monthly flows in all watersheds by Kolmogorov-Smirnov test while those distributions were judged to be unfitness in Nam Pyeong of Yeong San and Song Jeong and Ab Rog watersheds of Seom Jin river systems in the $\chi$$^2$ goodness of fit test. 2.Most of the arithmetic mean values for synthetic monthly flows simulated by Gamma distribution are much closer to the results of the observed data than those of two parameter lognomal distribution in the applied watersheds. 3.Fluctuation for the coefficient of variation derived by Gamma distribution was shown in general as better agreement with the results of the observed data than that of two parameter lognormal distribution in the applied watersheds both in Yeong San and Seom Jin river systems. Especially, coefficients of variation calculated by Gamma distribution are seemed to be much closer to those of the observed data during July and August. 4.It can be concluded that synthetic monthly flows simulated by Gamma distribution are seemed to be much closer to the observed data than those by two parameter lognormal distribution in the applied watersheds. 5.It is to be desired that multi-season first order Markov model based on Gamma distribution which is confirmed as a good fitting one in this study would be compared with Harmonic synthetic model as a continuation follows.

• Journal of the Korea Society for Simulation
• /
• v.19 no.4
• /
• pp.249-259
• /
• 2010

Synthetic environment data used in defense M&S fields, which came from various organization and source, are consumed and managed by their own native database system in distributed environment. But to manage these diverse data while interoperation in HLA/RTI environment, neutral synthetic environment data model is necessary to transmit the data between native database. By the support of DMSO, SEDRIS was developed to achieve this requirement and this specification guarantees loss-less data representation, interchange and interoperability. In this research, to use SEDRIS as a standard simulation database, base research, visualization for validation, data interchange experiment through test-bed was done. This paper shows each research case, result and future research direction, to propose standardized SEDRIS usage process.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.38 no.4
• /
• pp.110-124
• /
• 1996

This study was conducted to simulate long seres of synthetic monthly flows by multi-season first order Markov model with selection of best fitting frequency distribution, harmonic synthetic and harmonic regression models and to make a comparison of statistical parameters between observes and synthetic flows of five watersheds in Geum river system. The results obtained through this study can be summarized as follow. 1. Both gamma and two parameter lognormal distributions were found to be suitable ones for monthly flows in all watersheds by Kolmogorov-Smirnov test. 2. It was found that arithmetic mean values of synthetic monthly flows simulated by multi-season first order Markov model with gamma distribution are much closer to the results of the observed data in comparison with those of the other models in the applied watersheds. 3. The coefficients of variation, index of fluctuation for monthly flows simulated by multi-season first order Markov model with gamma distribution are appeared closer to those of the observed data in comparison with those of the other models in Geum river system. 4. Synthetic monthly flows were simulated over 100 years by multi-season first order Markov model with gamma distribution which is acknowledged as a suitable simulation modal in this study.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.22 no.1
• /
• pp.49-59
• /
• 2019

Generative adversarial networks(GANs) have received great attention in the machine learning field for their capacity to model high-dimensional and complex data distribution implicitly and generate new data samples from the model distribution. This paper investigates the model training methodology, architecture, and various applications of generative adversarial networks. Experimental evaluation is also conducted for generating synthetic image dataset for defense using two types of GANs. The first one is for military image generation utilizing the deep convolutional generative adversarial networks(DCGAN). The other is for visible-to-infrared image translation utilizing the cycle-consistent generative adversarial networks(CycleGAN). Each model can yield a great diversity of high-fidelity synthetic images compared to training ones. This result opens up the possibility of using inexpensive synthetic images for training neural networks while avoiding the enormous expense of collecting large amounts of hand-annotated real dataset.

• The Journal of Korea Robotics Society
• /
• v.14 no.2
• /
• pp.139-149
• /
• 2019

In this paper, we present auto-annotation tool and synthetic dataset using 3D CAD model for deep learning based object detection. To be used as training data for deep learning methods, class, segmentation, bounding-box, contour, and pose annotations of the object are needed. We propose an automated annotation tool and synthetic image generation. Our resulting synthetic dataset reflects occlusion between objects and applicable for both underwater and in-air environments. To verify our synthetic dataset, we use MASK R-CNN as a state-of-the-art method among object detection model using deep learning. For experiment, we make the experimental environment reflecting the actual underwater environment. We show that object detection model trained via our dataset show significantly accurate results and robustness for the underwater environment. Lastly, we verify that our synthetic dataset is suitable for deep learning model for the underwater environments.