• Korean Journal of Remote Sensing
• /
• v.40 no.4
• /
• pp.351-361
• /
• 2024

Satellite imagery contains various elements such as clouds, cloud shadows, and terrain shadows. Accurately identifying and eliminating these factors that complicate satellite image analysis is essential for maintaining the reliability of remote sensing imagery. For this reason, satellites such as Landsat-8, Sentinel-2, and Compact Advanced Satellite 500-1 (CAS500-1) provide Usable Data Masks(UDMs)with images as part of their Analysis Ready Data (ARD) product. Precise detection of clouds and their shadows is crucial for the accurate construction of these UDMs. Existing cloud and their shadow detection methods are categorized into threshold-based methods and Artificial Intelligence (AI)-based methods. Recently, AI-based methods, particularly deep learning networks, have been preferred due to their advantage in handling large datasets. This study aims to analyze the applicability of constructing UDMs for high-resolution satellite images through deep learning-based cloud and their shadow detection using open-source datasets. To validate the performance of the deep learning network, we compared the detection results generated by the network with pre-existing UDMs from Landsat-8, Sentinel-2, and CAS500-1 satellite images. The results demonstrated that high accuracy in the detection outcomes produced by the deep learning network. Additionally, we applied the network to detect cloud and their shadow in KOMPSAT-3/3A images, which do not provide UDMs. The experiment confirmed that the deep learning network effectively detected cloud and their shadow in high-resolution satellite images. Through this, we could demonstrate the applicability that UDM data for high-resolution satellite imagery can be constructed using the deep learning network.

• Journal of the Korea Computer Graphics Society
• /
• v.27 no.5
• /
• pp.45-54
• /
• 2021

Recently, deep learning-based denoising approaches have been actively studied. In particular, with the advances of blind denoising techniques, it become possible to train a deep learning-based denoising model only with noisy images in an image domain where it is impossible to obtain a clean image. We no longer require pairs of a clean image and a noisy image to obtain a restored clean image from the observation. However, it is difficult to recover the target using a deep learning-based denoising model trained by only noisy images if the distribution of the noisy image is far from the distribution of the clean image. To address this limitation, unpaired image denoising approaches have recently been studied that can learn the denoising model from unpaired data of the noisy image and the clean image. ISCL showed comparable performance close to that of supervised learning-based models based on pairs of clean and noisy images. In this study, we propose suitable normalization techniques for each purpose of architectures (e.g., generator, discriminator, and extractor) of ISCL. We demonstrate that the proposed method outperforms state-of-the-art unpaired image denoising approaches including ISCL.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.6
• /
• pp.420-433
• /
• 2023

With the development of ICT technologies, various systems and services based on data have been introduced. Also, raising the importance of technologies that provide trust to a data, infrastructure, and services. for ICT services. However, in the existing ICT service environment, there is no way to objectively judge and manage the trust of various components (content, infrastructure, process, service provider, etc) that constitute a specific service, and there is a limitation that we can only rely on the service provider's own quality standards. This paper provides requirements, functional archtecture, and procedures for providing reliable ICT services from the perspective of ITU-T international standards. Trust-enabled service provisioning adds additional functions for providing trust to existing ICT service entities (service resources, stakeholders, and users), collecting, analyzing, and providing trust related information. Therefore, a users can consider the trust of various service components based on analyzed trust information based on their trust requirements.

• Journal of KIISE
• /
• v.44 no.1
• /
• pp.21-26
• /
• 2017

In recent years, new technologies such as Internet of Things, Cloud Computing and Big Data are being widely used. And the type and amount of data is dramatically increasing. This makes security an important issue. In terms of leakage of sensitive personal information. In order to protect confidential information, a method called anonymization is used to remove personal identification elements or to substitute the data to some symbols before distributing and sharing the data. However, the existing method performs anonymization by generalizing the level of quasi-identifier hierarchical. It requires a higher level of generalization in case where k-anonymity is not satisfied since records in data table are either added or removed. Loss of information is inevitable from the process, which is one of the factors hindering the utility of data. In this paper, we propose a novel anonymization technique using decision tree based machine learning to improve the utility of data by minimizing the loss of information.

• Journal of KIISE
• /
• v.42 no.6
• /
• pp.817-822
• /
• 2015

Relation extraction is an important information extraction technique that can be widely used in areas such as question-answering and knowledge population. Previous studies on relation extraction have been based on supervised machine learning models that need a large amount of training data manually annotated with relation categories. Recently, to reduce the manual annotation efforts for constructing training data, distant supervision methods have been proposed. However, these methods suffer from a drawback: it is difficult to use these methods for collecting negative training data that are necessary for resolving classification problems. To overcome this drawback, we propose a one-class classification model that can be trained without using negative data. The proposed model determines whether an input data item is included in an inner category by using a similarity measure based on lexical information and syntactic patterns in a vector space. In the experiments conducted in this study, the proposed model showed higher performance (an F1-score of 0.6509 and an accuracy of 0.6833) than a representative one-class classification model, one-class SVM(Support Vector Machine).

• KIISE Transactions on Computing Practices
• /
• v.22 no.2
• /
• pp.101-106
• /
• 2016

A femtocell network has been attracting attention as a promising solution for providing high data rate transmission over the conventional cellular network in an indoor environment. In this paper, we propose a distributed power control algorithm considering both indoor coverage and data load balancing in the femtocell network. As data traffic varies by time and location according to user distribution, each femto base station suffers from an unbalanced data load, which may degrade network performance. To distribute the data load, the base stations are required to adjust their transmission power dynamically. Since there are a number of base stations in practice, we propose a distributed power control algorithm. In addition, we propose the simple algorithm to detect the faulty base station and to recover coverage. We also explain how to insert a new base station into a deployed network. We present the simulation results to evaluate the proposed algorithms.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.9
• /
• pp.1655-1665
• /
• 2017

Recently data centers are being constructed actively by many cloud service providers, enterprises, research institutes, etc. Generally, they are built on tree topology using ECMP data forwarding scheme for load balancing. In this paper, we examine data center network topologies like tree topology and fat-tree topology, and load balancing technologies like MLAG and ECMP. Then, we propose a scheduling algorithm to efficiently transmit particular files stored on the hosts in the data center to the destination node outside the data center, where fat-tree topology and OpenFlow protocol between infrastructure layer and control layer are used. We run performance analysis by numerical method, and compare the analysis results with those of ECMP. Through the performance comparison, we show the outperformance of the proposed algorithm in terms of throughput and file transfer completion time.

• Bulletin of the Korean Space Science Society
• /
• 2010.04a
• /
• pp.28.4-29
• /
• 2010

이 연구는 심우주 추적망(Deep Space Network) 측정 시스템의 구현을 위한 한국형 심우주 항법 예비 소프트웨어(Korean Deep Space Orbit Determination Program version 1; K-DSODP ver.1)의 개발을 목표로 한다. 연구의 주 내용은 심우주 항법을 위한 기초 기술 연구로 지구로부터 달까지 진행하는 탐사선의 궤적 추정에 대한 것이며, 연구의 시작에 앞서 사용될 관측 데이터를 얻기 위해 한국형 심우주 항법 관측데이터 생성 소프트웨어(Korean Deep Space Observation Data Generation Program version 1; K-DSODGP ver.1)를 개발하여 사용하였다. 임의의 잡음이 추가된 가상의 관측 데이터를 생성한 후, 이 관측 데이터를 실제 궤도로 상정하여 기하학적인 관측 모델을 수립하였고, 일정한 시간 간격동안 모은 임의의 관측 데이터를 가지고 궤도 결정을 수행하여 추정된 궤도를 전파하였다. 궤도 결정 알고리즘을 구성하기 위해 기본적인 좌표계, 탐사선에 미치는 지구의 중력에 대한 동역학 모델, 천체력과 탐사선의 동역학 모델로 구성된 관측 모델들을 유도하였으며, 탐사선의 위치와 속도를 추정하는 과정에서 가중치 최소 자승법을 적용하여 추정 궤도와 실제 궤도의 최소화를 유도하였다. 이러한 일련의 과정을 통해 요구한 시각의 탐사선의 위치와 속도를 결정하는 궤도결정 시스템을 구현하였고, 궤도 결정 시스템의 성능을 평가하기 위해 전파된 궤도와 실제 궤도의 차이를 분석하였다. 결과적으로 300초마다 관측데이터를 받을 경우, 2일 이상의 궤도결정 시간간격을 상정했을 때 평균 오차는 각각 약 0.26km RMS(range), 6.84km/s RMS(range-rate) 이내의 결과를 얻었고, 600초마다 관측데이터를 받을 경우, 평균 오차는 각각 약 0.30km RMS (range), 6.35km/s RMS(range-rate) 이내의 안정적인 결과를 얻었다. 이 연구의 결과를 통하여 추후 심화된 심우주 항법 소프트웨어 개발을 위한 기반이 마련될 것이다.

• Journal of KIISE
• /
• v.42 no.5
• /
• pp.580-590
• /
• 2015

Recently, the issue of efficient power management in datacenters as a part of green computing has gained prominence. For realizing efficient power management, effective power monitoring and analysis must be conducted for servers in a datacenter. However, an administrator should know the exact structure of the datacenter and its associated databases, and is required to analyze relationships among the observed data. This is because according to previous monitoring approaches, servers are usually managed using only databases. In addition, it is not possible to monitor data that are not indicated in databases. To overcome these drawbacks, we proposed an ontology-based monitoring approach. We constructed domain ontology for management servers at a datacenter, and mapped observed data onto the constructed domain ontology by using semantic annotation. Moreover, we defined query creation rules and server state rules. To demonstrate the proposed approach, we designed an ontology based monitoring system architecture, and constructed a knowledge system. Subsequently, we implemented the pilot system to verify its effectiveness.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.32 no.3
• /
• pp.23-44
• /
• 2021

This research proposed an approach for interlinking bibliographic data used in the existing multicultural institutions in order to support information services of multicultural institutions. It constructed a set of linkage properties that can link the FRBR-based BIBFRAME structure with the existing metadata standards currently adopted in multicultural institutions, including MODS, EAD, and CDWA. As a result, the descriptive elements for the Work level, Instance level, Item level, statement of responsibility, subjects were established in the BIBFRAME syntax. A total of 19 upper level elements, 18 sub-elements, and 12 linkage properties were proposed for the linking of bibliographic data, so that a bibliographic environment was established that can interconnect heterogeneous metadata adopted depending on the types of collections. Through the proposed structure, it is expected that it can be applied to provide sufficient information services more efficiently by linking the dispersed bibliographic data of multicultural institutions in an integrated way.