• Journal of the Korea Institute of Military Science and Technology
• /
• v.27 no.3
• /
• pp.319-328
• /
• 2024

Semi-supervised learning is a good way to train a classification model using a small number of labeled and large number of unlabeled data. We applied semi-supervised learning to a synthetic aperture radar(SAR) image classification model with a limited number of datasets that are difficult to create. To address the previous difficulties, semi-supervised learning uses a model trained with a small amount of labeled data to generate and learn pseudo labels. Besides, a lot of number of papers use a single fixed threshold to create pseudo labels. In this paper, we present a semi-supervised synthetic aperture radar(SAR) image classification method that applies different thresholds for each class instead of all classes sharing a fixed threshold to improve SAR classification performance with a small number of labeled datasets.

• Journal of information and communication convergence engineering
• /
• v.21 no.1
• /
• pp.98-102
• /
• 2023

This paper presents a study on how augmenting semi-synthetic image data improves the performance of human detection algorithms. In the field of object detection, securing a high-quality data set plays the most important role in training deep learning algorithms. Recently, the acquisition of real image data has become time consuming and expensive; therefore, research using synthesized data has been conducted. Synthetic data haves the advantage of being able to generate a vast amount of data and accurately label it. However, the utility of synthetic data in human detection has not yet been demonstrated. Therefore, we use You Only Look Once (YOLO), the object detection algorithm most commonly used, to experimentally analyze the effect of synthetic data augmentation on human detection performance. As a result of training YOLO using the Penn-Fudan dataset, it was shown that the YOLO network model trained on a dataset augmented with synthetic data provided high-performance results in terms of the Precision-Recall Curve and F1-Confidence Curve.

• Journal of the Korean Data and Information Science Society
• /
• v.23 no.3
• /
• pp.579-585
• /
• 2012

Many different semi-supervised learning algorithms have been proposed for use wit unlabeled data. However, most of them focus on classification problems. In this paper we propose a semi-supervised regression algorithm called the semi-supervised local constant estimator (SSLCE), based on the local constant estimator (LCE), and reveal the asymptotic properties of SSLCE. We also show that the SSLCE has a faster convergence rate than that of the LCE when a well chosen weighting factor is employed. Our experiment with synthetic data shows that the SSLCE can improve performance with unlabeled data, and we recommend its use with the proper size of unlabeled data.

• Korean Journal of Remote Sensing
• /
• v.33 no.4
• /
• pp.437-444
• /
• 2017

Since land pixels often generate false alarms in ship detection using Synthetic Aperture Radar (SAR), land masking is a necessary step which can be processed by a land area map or water database. However, due to the continuous coastline changes caused by newport, bridge, etc., an updated data should be considered to mask either the land or the oceanic part of SAR. Furthermore, coastal concrete facilities make noise signals, mainly caused by side lobe effect. In this paper, we propose two methods. One is a semi-automatic water body data generation method that consists of terrain correction, thresholding, and median filter. Another is a dynamic land masking method based on water database. Based on water database, it uses a breadth-first search algorithm to find and mask noise signals from coastal concrete facilities. We verified our methods using Sentinel-1 SAR data. The result shows that proposed methods remove maximum 84.42% of false alarms.

• Journal of Korea Game Society
• /
• v.13 no.2
• /
• pp.111-118
• /
• 2013

The US military is developing and improving SE-CORE and CVE to integrate war game between different kinds of systems, establish time-optimal CVE. Although South Korea is researching SEDRIS, etc. that is suitable for domestic circumstances, there are still many problems to be solved. The researchers suggest a way to semi-automate part of SNE process, by using the channel information of the image. This can help to improve speed of creating synthetic environment and facilitate information sharing to the others. If there is further research on the subject, it will be possible to develop automation technology to apply data from various information collection devices to synthetic battlefield environment (SBE).

• International Journal of Computer Science & Network Security
• /
• v.22 no.6
• /
• pp.230-240
• /
• 2022

Sharing of online videos via internet is an emerging and important concept in different types of applications like surveillance and video mobile search in different web related applications. So there is need to manage personalized web video retrieval system necessary to explore relevant videos and it helps to peoples who are searching for efficient video relates to specific big data content. To evaluate this process, attributes/features with reduction of dimensionality are computed from videos to explore discriminative aspects of scene in video based on shape, histogram, and texture, annotation of object, co-ordination, color and contour data. Dimensionality reduction is mainly depends on extraction of feature and selection of feature in multi labeled data retrieval from multimedia related data. Many of the researchers are implemented different techniques/approaches to reduce dimensionality based on visual features of video data. But all the techniques have disadvantages and advantages in reduction of dimensionality with advanced features in video retrieval. In this research, we present a Novel Intent based Dimension Reduction Semi-Supervised Learning Approach (NIDRSLA) that examine the reduction of dimensionality with explore exact and fast video retrieval based on different visual features. For dimensionality reduction, NIDRSLA learns the matrix of projection by increasing the dependence between enlarged data and projected space features. Proposed approach also addressed the aforementioned issue (i.e. Segmentation of video with frame selection using low level features and high level features) with efficient object annotation for video representation. Experiments performed on synthetic data set, it demonstrate the efficiency of proposed approach with traditional state-of-the-art video retrieval methodologies.

• BMB Reports
• /
• v.31 no.6
• /
• pp.525-535
• /
• 1998

A critical step in the faithful translation of genetic information is specific tRNA recognition by aminoacyl-tRNA synthetases. These enzymes catalyze the covalent attachment of particular amino acids to the terminal adenosine of cognate tRNA substrates. In general, there is one synthetase for each of the twenty amino acids and each enzyme must discriminate against all of the cellular tRNAs that are specific for the nineteen noncognate amino acids. Primary sequence information combined with structural data have resulted in the division of the twenty synthetases into two classes. In recent years, several high-resolution co-crystal structures along with biochemical data have led to an increased understanding of tRNA recognition by synthetases of both classes. The anticodon sequence and the amino acid acceptor stem are the most common locations for critical recognition elements. This review will focus on acceptor stem discrimination by class II synthetases. In particular, the results of in vitro aminoacylation assays and site-directed and atomic group mutagenesis studies will be discussed. These studies have revealed that even subtle atomic determinants can provide signals for specific tRNA aminoacylation.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.152-152
• /
• 2022

Only employing historical data limits the estimation of the full distribution of probable Tropical Cyclone (TC) risk due to the insufficiency of samples. Addressing this limitation, this study introduces a semi-physical TC rainfall model that produces spatially and temporally resolved TC rainfall data to improve TC risk assessments. The model combines a statistical-based track model based on the Markov renewal process to produce synthetic TC tracks, with a physics-based model that considers the interaction between TC and the atmospheric environment to estimate TC rainfall. The simulated data from the combined model are then fitted to a probability distribution function to compute the spatially heterogeneous risk brought by landfalling TCs. The methodology is employed in South Korea as a case study to be able to implement a country-scale-based vulnerability inspection from damaging TC impacts. Results show that the proposed model can produce TC tracks that do not only follow the spatial distribution of past TCs but also reveal new paths that could be utilized to consider events outside of what has been historically observed. The model is also found to be suitable for properly estimating the total rainfall induced by landfalling TCs across various points of interest within the study area. The simulated TC rainfall data enable us to reliably estimate extreme rainfall from higher return periods that are often overlooked when only the historical data is employed. In addition, the model can properly describe the distribution of rainfall extremes that show a heterogeneous pattern throughout the study area and that vary per return period. Overall, results show that the proposed approach can be a valuable tool in providing sufficient TC rainfall samples that could be an aid in improving TC risk assessment.

• Journal of electromagnetic engineering and science
• /
• v.5 no.4
• /
• pp.183-188
• /
• 2005

A simple microwave backscattering model for vegetation canopies on earth surfaces is developed in this study. A natural earth surface is modeled as a two-layer structure comprising a vegetation layer and a ground layer. This scattering model includes various scattering mechanisms up to the first-order multiple scattering( double-bounce scattering). Radar backscatter from ground surface has been modeled by the polarimetric semi-empirical model (PSEM), while the backscatter from the vegetation layer modeled by the vector radiative transfer model. The vegetation layer is modeled by random distribution of mixed scattering particles, such as leaves, branches and trunks. The number of input parameters has been minimized to simplify the scattering model. The computation results are compared with the experimental measurements, which were obtained by ground-based scatterometers and NASA/JPL air-borne synthetic aperture radar(SAR) system. It was found that the scattering model agrees well with the experimental data, even though the model used only ten input parameters.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_3
• /
• pp.1761-1775
• /
• 2022

Due to unpredictable climate change, the frequency of occurrence of water-related disasters and the scale of damage are also continuously increasing. In terms of disaster management, it is essential to identify the damaged area in a wide area and monitor for mid-term and long-term forecasting. In the field of water disasters, research on remote sensing technology using Synthetic Aperture Radar (SAR) satellite images for wide-area monitoring is being actively conducted. Time-series analysis for monitoring requires a complex preprocessing process that collects a large amount of images and considers the noisy radar characteristics, and for this, a considerable amount of time is required. With the recent development of cloud computing technology, many platforms capable of performing spatiotemporal analysis using satellite big data have been proposed. Google Earth Engine (GEE)is a representative platform that provides about 600 satellite data for free and enables semi real time space time analysis based on the analysis preparation data of satellite images. Therefore, in this study, immediate water disaster damage detection and mid to long term time series observation studies were conducted using GEE. Through the Otsu technique, which is mainly used for change detection, changes in river width and flood area due to river flooding were confirmed, centered on the torrential rains that occurred in 2020. In addition, in terms of disaster management, the change trend of the time series waterbody from 2018 to 2022 was confirmed. The short processing time through javascript based coding, and the strength of spatiotemporal analysis and result expression, are expected to enable use in the field of water disasters. In addition, it is expected that the field of application will be expanded through connection with various satellite bigdata in the future.