최근 딥러닝 기술 가운데 이미지데이타 분석에 뛰어난 성능을 보이는 합성곱신경망 기술의 발전은 이미지 분석 영역에서 다양한 가능성을 제시하고 있다. 관광객이 게시한 사진을 딥러닝 기술을 이용하여 분류하기 위해서는 관광사진에 대한 분류와 목적에 맞는 딥러닝 모델의 훈련작업이 필수적으로 선행되어야 한다. 본 연구에서는 관광객이 플리커에 게시한 사진을 효율적으로 분류하기 위해 관광목적으로 사진이 어떻게 분류되어야 하는지 관광목적 사진분류 체계를 개발하고자 하였다. 관광목적 사진 분류 카테고리 개발을 위해 문헌분석, 웹사이트 분석, 관광객이 게시한 약 38,000장 사진의 검토과정을 거쳐 사진 분류 카테고리를 개발하였으며, 약 8400장의 사진을 개발된 카테고리에 맞춰 분류해 봄으로써 개발된 카테고리의 검증과정을 거쳤다. 이 과정을 거쳐 최종으로 제안된 카테고리는 13개 대분류, 64개 중분류, 164개의 세분류 체계를 갖으며, 본 연구 결과는 향후 관광목적 사진을 딥러닝 모델을 이용하여 분류하고자 할 때 기초자료로 활용될 것으로 기대된다.

In order to implement the smart home environment, we need an intelligence service platform that learns the user’s life style and behavioral patterns, and recommends appropriate services to the user. The intelligence service platform should embed a couple of effective and efficient data mining algorithms for learning from the data that is gathered from the smart home environment. In this study, we evaluate the suitability of data mining algorithms for smart home intelligent service platforms. In order to do this, we first develop an intelligent service scenario for smart home environment, which is utilized to derive functional and technical requirements for data mining algorithms that is equipped in the smart home intelligent service platform. We then evaluate the suitability of several data mining algorithms by employing the analytic hierarchy process technique. Applying the analytical hierarchy process technique, we first score the importance of functional and technical requirements through a hierarchical structure of pairwise comparisons made by experts, and then assess the suitability of data mining algorithms for each functional and technical requirements. There are several studies for smart home service and platforms, but most of the study have focused on a certain smart home service or a certain service platform implementation. In this study, we focus on the general requirements and suitability of data mining algorithms themselves that are equipped in smart home intelligent service platform. As a result, we provide a general guideline to choose appropriate data mining techniques when building a smart home intelligent service platform.

In cloud computing environment, service contracts between service providers and customers should be monitored in real-time. In the current paper-based service contract environment, service providers and customers cannot monitor service levels in real-time. Moreover, if a service failure occurs, there is no compensation system for the failure. Therefore, it is necessary to develop an automated administrative tool to support overall cloud services. The proposed intelligent framework is a document containing information processing knowledge which is used whenever information-processing demand occurs. It can be an important technology for service extension.

To solve the traffic problems of small cities like Chung Ju City, electronic communications equipment and the introduction of advanced technologies are being used or promoted. Other local areas with developed systems in place are being investigated for applicability. Systems like these will be introduced. This system will embody an organized, efficient transportation and safety system within this study At the moment, to solve the problem of reoccurring traffic in small towns in our country, we must develop an ITS service. We are now in the process of selecting and establishing the most adaptable ITS service to a small city For this purpose, we have studied existing records; taken a poll of the citizens of Chung Ju; examined present traffic conditions, examined road capacities; and reviewed available means of transportation. To accomplish the purpose of a thesis like this, we will research and analyze developed nations' and domestic cities' existing systems, construction and operation. We are in the process of building an ITS System which is the most adaptable to a small city like Chung Ju.

정보통신기술의 발달과 함께 학습자에게 재미와 흥미를 통한 즐거운 교육환경을 제공하고자 다양한 방법이 시도 되고 있다. 에듀테인먼트와 게임기반학습 등에서 게임이나 로봇과 같은 기술을 교육에 활용하는 것은 좋은 예이다. 본 연구에서는 휴머노이드 로봇의 율동 생성을 위한 사용자 데이터 수집과 분석을 통한 지능형 율동 교육 시스템을 제안한다. 이를 위하여 사용자는 음악을 선택하고, 선택한 음악에 따라 율동 정보를 입력한다. 이러한 사용자의 로봇 활용 데이터는 분석을 통하여 지능화된 서비스를 위한 패턴의 역할을 한다. 분석 결과는 빈도에 기반을 두며, 과거 정보가 부족한 경우 FFT 유사도 비교 방법을 적용하였다. 제안하는 방법은 유치원 아이들을 대상으로 하는 실험을 통하여 유효함을 확인하였다.

This study developed the prototype of the system and implemented its main functions, which is the intelligent integrated agricultural water management information system and service (IaWAMISS). The developed system was designed to be able to collect, process and analyze the agricultural water information of spatially dispersed reservoirs in whole country and spatial geographic information distributed in various systems of other organizations. The system, IaWAMISS, is also possible to provide the reproduced information services in each reservoir and space units, such as agricultural water demand and supply analysis and drought prediction, to the people, experts, and policy makers. This study defined the 6 step modules to develop the system, which are to design the components of intelligent integrated information system, to derive the utilization contents of existing systems, to design the new development elements for IaWAMISS, to design the reservoir information system can be used by managers of city and county, to designate the monitoring reservoirs managed by city and county, and finally to prepare the sharing system between organizations with the existing information systems. In order to implement the prototype of the system, this study shows the results for three important functions of the system: spatial integration of reservoirs' information, data link integration between the existing systems, and intelligent analysis program development to assist decision support for agricultural water management. For the spatial integration with the reservoir water information of the Korea Rural Community Corporation, this study get IaWAMISS to receive the real-time reservoir storage information from the measurement facility installed in the municipal management reservoir. The data link integration connecting databases of the existing systems, was implemented by integrating the meteorological information of the Korea Meteorological Administration with IaWAMISS, so that the rainfall forecast data could be derived and used. For the implementation of the intelligent analysis program, this study also showed the results of analysis and prediction of agricultural water demand and supply amount, estimation of Palmer drought index, analysis of flood risk area in typhoon course region, and analysis of the storage status of reservoirs related to each storm. This study confirmed the possibility and efficiency of an useful system development through the prototype design and implementation of IaWAMISS. By solving the preliminary 6 step modules presented in this study, it is possible not only to efficiently manage water by spatial unit, but also to provide the service of information and to enhance the relevant policy and national understanding to the people.

In this paper, we discuss a service to give alarm in the case of emergency for the old and the weak by human behavior recognition in Intelligent Space. Our Intelligent Space consists of mobile robots, sensors and agents. Agent analyzes data acquired from networked devices and determines tasks of robot and space to provide a service for humans. In our emergency alarm service, human behavior recognition service module analyzes accelerometer data obtained from body-attached human behavior sensing platform, and classifies into four basic human behavior such as "walking", "running", "sitting" and "falling-down". For the old and the weak, "falling-down" behavior may bring about dangerous situations. On such an occasion, emergency alarm service is executed automatically and a selected mobile robot approaches fallen person quickly and then send images of the person to guardians. In this paper, we set up a scenario to verify the emergency alarm service in Intelligent Space, and show feasibility of the service from our simulation experiments.

지능형 서비스 로봇은 인공지능 및 센서인터페이스 등의 다양한 기능과 IT기술의 접목으로 우리의 생활을 윤택하게 해줄 수 있는 인간 지향적인 성격으로 발전하고 있다. 또한 많은 분야에서 새로운 서비스 개발을 통해서 좀더 인간의 생활에 가까이 접근되어 지고 있다. 한편 국내 초등학교에서 초등교사는 학생 수의 과다로 능동적이기 보다는 수동적인 수업을 하고 있어서 학습 효율성이나 능률성이 떨어지고 있다. 이에 본 연구에서는 발전하고 있는 서비스 로봇 기술로 능