• Korean Journal of Remote Sensing
• /
• v.33 no.4
• /
• pp.423-436
• /
• 2017

This paper presents a deep learning-based road segmentation framework from very high-resolution orthophotos. The proposed method uses Deep Convolutional Autoencoders for end-to-end mapping of orthophotos to road segmentations. In addition, a set of post-processing steps were applied to make the model outputs GIS-ready data that could be useful for various applications. The optimization of the model's parameters is explained which was conducted via grid search method. The model was trained and implemented in Keras, a high-level deep learning framework run on top of Tensorflow. The results show that the proposed model with the best-obtained hyperparameters could segment road objects from orthophotos at an average accuracy of 88.5%. The results of optimization revealed that the best optimization algorithm and activation function for the studied task are Stochastic Gradient Descent (SGD) and Exponential Linear Unit (ELU), respectively. In addition, the best numbers of convolutional filters were found to be 8 for the first and second layers and 128 for the third and fourth layers of the proposed network architecture. Moreover, the analysis on the time complexity of the model showed that the model could be trained in 4 hours and 50 minutes on 1024 high-resolution images of size $106{\times}106pixels$, and segment road objects from similar size and resolution images in around 14 minutes. The results show that the deep learning models such as Convolutional Autoencoders could be a best alternative to traditional machine learning models for road segmentation from aerial photographs.

• Journal of the Korean Society for information Management
• /
• v.35 no.1
• /
• pp.129-155
• /
• 2018

This research aims to discover various aspects of the user studies and the research in practice and also to propose collaboration methods by empirical analysis of the data. To determine the application applicability of the user studies in other subject areas, the degree of keyword overlap between the user studies and the User Experience (UX), one of the research in practice discipline, was measured. The quantitative information science methods including simple frequency analysis were applied to more than ten thousand published papers to generate the network mapping and ranking as well as comparative analysis by time. The analysis result showed that there were slightly lesser overlap between the user studies and the UX in the domestically published articles than the international ones. It also revealed that there is a relationship between the actual occurrences of collaboration and the keyword overlap. The temporal analysis showed that there is increasingly more keyword overlap between two disciplines and thus it is possible to predict the active convergence in the future.

• Journal of Positioning, Navigation, and Timing
• /
• v.9 no.3
• /
• pp.229-236
• /
• 2020

The trend of water vapor contents in atmosphere is one of key elements for studying climate change. The tropospheric products, i.e., ZTD values achieved through GPS data processing can retrieve the amount of water vapor with higher temporal and spatial resolution than any other instruments. In this study, the tropospheric products of KASINET for a time period from 2001 to 2014 are reprocessed using PPP strategy and the products from the CODE's 2nd reprocessing campaign. For consistency with reprocessing activities of other networks like EPN, the VMF1 mapping function and non-tidal loading effect due to atmospheric pressure are applied in the process. The reprocessing results are investigated through comparing with the CODE's 2nd reprocessing products by including some IGS stations in the process and also calculating weekly coordinate repeatability to see the quality of the processing. After removing outliers based on the variation of averaged formal error, all processed stations have similar variations of formal error about 2 mm which is lower than that of the IGS final product. Comparison results with the CODE's 2nd reprocessing products show that the overall mean difference is found to be -0.28±5.54 mm which is similar level of the previous studies. Finally, the ZTD trends of all KASINET stations are calculated and the averaged trend is achieved as 0.19 mm/yr. However, the trend of each month shows different amounts and directions from -1.26 mm/yr in May to 1.18 mm/yr in August. In conclusion, the reprocessed tropospheric product and applied strategy of this study has enough quality as one of reliable solution for a reference product for Korean Peninsula which is needed to use GPSbased tropospheric product for climate change research.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.3
• /
• pp.73-81
• /
• 2020

Today's massively multiplayer online games(MMOGs) can contain millions of synchronous players scattered across the world and participating with each other within a single shared game. The increase in the number of players in MMOGs has led to some issues with the demand of server which generates a significant increase in costs for the game industry and impacts to the quality of service offered to players. In dealing with a considerable scale of MMOGs, we propose a cloud computing and peer-to-peer(P2P) hybrid architecture in this paper. Given the two nearly independent functionalities of P2P and cloud architectures, we consider the possibility of fusing these two concepts and researching the application of the resultant amalgamation in MMOGs. With an efficient and effective provisioning of resources and mapping of load, the proposed hybrid architecture relieves a lot of computational power and network traffic, the load on the servers in the cloud while exploiting the capacity of the peers. The simulation results show that MMOGs based on the proposed hybrid architecture have better performance and lower traffic received compared with MMOGs based on traditional client-server system.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.9 no.2
• /
• pp.11-19
• /
• 2006

Sewerage is the system that improves the quality of human life and prevents many disasters such as floods. However the investigators in Korea only have been concerned about the sewer system, so the sewage treatment plant stays in the basic level like mapping. For example, only one attribute can be recognized in the linear object. Because of this limitation, it makes difficult to manage the linear attribute regarding to the sewage pipe plan. And it is impossible to control a partial (point type, line type) attribute changes of the linear object. We will therefore present the applicable method for the attribute changes of the linear object like the sewage pipe plans. For this reason, this paper is designed on the basis of Dynamic Segmentation(DS). DS has the advantage of giving the attribute value to the exact place in the linear object. As a result of using DS, the variety environment changes around the sewage pipes are applied to the building sewerage data. This also makes it possible to get a precise estimation for the maximum dirty water amount.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.12 no.3
• /
• pp.99-114
• /
• 2012

Proxy Mobile IPv6 (PMIPv6) has received considerable attention between telecommunications and the Internet communities and does not require active participation of the Mobile Node (MN) by way of network-based mobility management. The PMIPv6 domain is studying establishment in progress to support extensively a number of MN by using a low handover latency. In this research, we are propose a novel 3S scheme for building Scalable and Secure and Seamless PMIPv6 domains. In the proposed scheme, all of Mobility Access Gateway (MAG) are acting as the Local Mobility Anchor (LMA) and composing a virtual ring with another MAG. General hashing is used in the efficient distribution-mapping between each MN and the MN's LMA of all MAGs. And, MAG and MN are authenticated using the symmetric key. Through mathematical analysis, we verifies the safety, scalability, and seamless service for 3S. Also, we're propose a handover procedure of 3S and show better than the existing schemes in terms of handover latency.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.6 no.2
• /
• pp.257-263
• /
• 2002

A graph embedding problem has been studied for applications of resource allocation and mapping the underlying data structure of a parallel algorithm into the interconnection architecture of massively parallel processing systems. In this paper, we consider the embedding problem of the pyramid into the regular 2-dimensional mesh interconnection network topology. We propose a new embedding function which can embed the pyramid of height N into 2$^{N}$ x2$^{N}$ 2-dimensional mesh with dilation max{2$^{N1}$-2. [3.2$^{N4}$＋1)/2, 2$^{N3}$＋2. [3.2$^{N4}$＋1)/2]}. This means an improvement in the dilation measure from 2$^{N}$ $^1$in the previous result into about (5/8) . 2$^{N1}$ under the same condition.condition.

• Spatial Information Research
• /
• v.6 no.1
• /
• pp.65-76
• /
• 1998

Air contaminant density must be inferred exactly to manage air pollution. Each land use of air pollution source is duplicated in the existing air contaminant distribution because the resolution of the land use map is low. The purpose of this study is to understand how the land use map is used to determine effectively in the distribution calculation of the emission volume and the inference of air contaminant density, as it is made in a high resolution. The major findings are as follows : In this study, as to making a high resolution($28.5m{\times}28.5m$) map of land use with GIS, each air pollution source is not duplicated spatially and land use can be reflected effectively. In Seoul, each air contaminant density was inferred (using a TCM-2 model) with the existing distribution map of emission volume, whose resolution is $1km{\times}1km$, and the new distribution map of emission volume, whose resolution is $28.5km{\times}28.5km$. According to the result, the inference value of the new distribution map was more similar to the actual value of an automatic survey network.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2001.10a
• /
• pp.165-196
• /
• 2001

DNA chip 또는 microarray는 다수의 유전자 또는 유전자 조각을 (보통 수천내지 수만 개)칩상에 고정시켜 놓고 DNA hybridization 반응을 이용하여 유전자들의 발현 양상을 분석할 수 있는 기술이다. 이러한 high-throughput기술은 예전에는 생각하지 못했던 여러가지 분자생물학의 문제에 대한 해답을 제시해 줄 수 있을 뿐 만 아니라, 분자수준에서의 질병 진단, 신약 개발, 환경 오염 문제의 해결 등 그 응용 가능성이 무한하다. 이 기술의 실용적인 적용을 위해서는 DNA chip을 제작하기 위한 하드웨어/웻웨어 기술 외에도 이러한 데이터로부터 최대한 유용하고 새로운 지식을 창출하기 위한 bioinformatics 기술이 핵심이라고 할 수 있다. 유전자 발현 패턴을 데이터마이닝하는 문제는 크게 clustering, classification, dependency analysis로 구분할 수 있으며 이러한 기술은 통계학과인공지능 기계학습에 기반을 두고 있다. 주로 사용된 기법으로는 principal component analysis, hierarchical clustering, k-means, self-organizing maps, decision trees, multilayer perceptron neural networks, association rules 등이다. 본 세미나에서는 이러한 기본적인 기계학습 기술 외에 최근에 연구되고 있는 새로운 학습 기술로서 probabilistic graphical model (PGM)을 소개하고 이를 DNA chip 데이터 분석에 응용하는 연구를 살펴본다. PGM은 인공신경망, 그래프 이론, 확률 이론이 결합되어 형성된 기계학습 모델로서 인간 두뇌의 기억과 학습 기작에 기반을 두고 있으며 다른 기계학습 모델과의 큰 차이점 중의 하나는 generative model이라는 것이다. 즉 일단 모델이 만들어지면 이것으로부터 새로운 데이터를 생성할 수 있는 능력이 있어서, 만들어진 모델을 검증하고 이로부터 새로운 사실을 추론해 낼 수 있어 biological data mining 문제에서와 같이 새로운 지식을 발견하는 exploratory analysis에 적합하다. 또한probabilistic graphical model은 기존의 신경망 모델과는 달리 deterministic한의사결정이 아니라 확률에 기반한 soft inference를 하고 학습된 모델로부터 관련된 요인들간의 인과관계(causal relationship) 또는 상호의존관계(dependency)를 분석하기에 적합한 장점이 있다. 군체적인 PGM 모델의 예로서, Bayesian network, nonnegative matrix factorization (NMF), generative topographic mapping (GTM)의 구조와 학습 및 추론알고리즘을소개하고 이를 DNA칩 데이터 분석 평가 대회인 CAMDA-2000과 CAMDA-2001에서 사용된cancer diagnosis 문제와 gene-drug dependency analysis 문제에 적용한 결과를 살펴본다.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.11 no.6
• /
• pp.73-80
• /
• 2011

In this paper, we design and implementation of effective remote management system for VoIP(Voice over IP) terminal. Ordinary VoIP terminal is connected inside of NAT(Network Address Translation) assigned private IP address. NAT provides address mapping between outside public IP address and inside private IP address, for outside system telephone call is connected to inside VoIP terminal. In this paper, we also design and implementation of UDP hole punching function that the outside calls pass through the NAT, for outside management system call is connected to internal VoIP terminal, and gathered to management information of the VoIP terminal. Also, we evaluated and analysed of developed system in the test environment. As a result of test, It showed that it was well performed without any data error and data loss in the connectivity, and It showed that it was well gathered management information of the VoIP terminal.