• Journal of Intelligence and Information Systems
• /
• v.28 no.2
• /
• pp.101-125
• /
• 2022

Recently, with the development of computing technology and the improvement of the cloud environment, deep learning technology has developed, and attempts to apply deep learning to various fields are increasing. A typical example is anomaly detection, which is a technique for identifying values or patterns that deviate from normal data. Among the representative types of anomaly detection, it is very difficult to detect a contextual anomaly that requires understanding of the overall situation. In general, detection of anomalies in image data is performed using a pre-trained model trained on large data. However, since this pre-trained model was created by focusing on object classification of images, there is a limit to be applied to anomaly detection that needs to understand complex situations created by various objects. Therefore, in this study, we newly propose a two-step pre-trained model for detecting abnormal situation. Our methodology performs additional learning from image captioning to understand not only mere objects but also the complicated situation created by them. Specifically, the proposed methodology transfers knowledge of the pre-trained model that has learned object classification with ImageNet data to the image captioning model, and uses the caption that describes the situation represented by the image. Afterwards, the weight obtained by learning the situational characteristics through images and captions is extracted and fine-tuning is performed to generate an anomaly detection model. To evaluate the performance of the proposed methodology, an anomaly detection experiment was performed on 400 situational images and the experimental results showed that the proposed methodology was superior in terms of anomaly detection accuracy and F1-score compared to the existing traditional pre-trained model.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.31 no.6
• /
• pp.282-286
• /
• 2021

Penetration depth and compressive residual stress of laser-carburized TiZrN coating by thickness of carbon paste were investigated in terms of carbon potential. The carbon paste was covered with a thickness of 1.1 mm using screen printing, and applied to a thickness of 0.4 mm using spin coating, and laser carburization was performed under the same conditions. As the thickness of carbon paste increased, the diffraction pattern of the laser-carburized TiZrN coating shifted to a lower angle, indicating solid solution strengthening and lattice distortion. For microstructure analysis using TEM, the defects and carbon concentration of the laser-carburized TiZrN coating increased as the carbon paste was thicker. It indicated that the variation of the carbon potential corresponds to the change in the paste thickness. In XPS depth profile analysis, high concentration of carbon and formation of carbide were observed in laser-carburized TiZrN coating with thick carbon paste. It revealed that the carbon concentration on the surface and carbon potential were changed by the thickness control of carbon paste. The compressive residual stress increased from 3.67 GPa to 4.58 GPa by the variation of carbon concentration.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_3
• /
• pp.863-872
• /
• 2022

Rare earth elements (REEs) have been used as one of power tracers for understanding geological and environmental changes due to their similar physico-chemical properties. In this study, we investigated the characteristics of rare earth elements in atmospheric particulates(PM_2.5) collected in Beijing and Gwangju during January 2018. The total concentrations of REEs in the Beijing samples were about 16X higher than those in Gwangju samples, and both samples are enriched in light REE than heavy REE, up to 8-10 times. The Post Archean Australian Shale (PAAS)-normalized pattern showed that both samples are enriched in Eu, Tb, and Er, and displayed positive Eu but negative Ce anomalies. The elemental correlations indicate that both samples originated from China desert and Loess plateau as well as cities surrounding Beijing. This study suggests that REEs in PM_2.5 can be used as a powerful proxy of revealing the difference between China and Korea, and provide basic information on the source and transport of PM_2.5.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.31-31
• /
• 2022

기후변화, 물 부족, 인구 증가와 도시화로 인한 물 수요 증가, 수질 악화, 노후화된 인프라와 같은 세계적인 물문제의 증가로 인해, 도시 물순환 시스템 관리는 더 큰 어려움을 겪고 있다. 취수, 도·송수, 정수처리, 배·급수, 용수 사용, 하수 집수, 하수 처리, 재이용 및 배출 과정을 포함하는 도시 물순환 시스템의 과정은 매우 에너지 집약적인 활동이며, 이와 같은 에너지 소비는 탄소 배출과 양의 직접적인 상관관계가 있다. 따라서 자원 관리 및 데이터 관리를 최적화하기 위해 넥서스 관점의 접근법이 도시 물순환 시스템에 점차적으로 도입되고 있는 추세이다. 도시 물순환 시스템 넥서스에서는 일반적으로 에너지 인텐시티로 표현되는 물을 위한 에너지를 이해하는 것이 중요하다. 에너지 인텐시티의 차이는 기후(연평균 강수량, 단기 기후 변동성, 기후패턴 등), 지리적 특징(표고차, 평지비, 위치 등), 시스템 특성(총급수량, 인구, 인구밀도, 관로 연장 등) 및 운영관리 효율(수압, 누수율, 에너지 효율 등)과 밀접한 관계가 있다. 그리고 도시 물순환 시스템에서 에너지 관리를 증진시킨 방안은 유지관리 효율 개선(물·에너지 관리전략, 물손실 관리, 수요 관리 및 수요 대응 등), 신기술 도입, 그리고 에너지 회수로 나누어진다. 본 연구에서는 기존 문헌의 자료를 분석하여 도시 물순환 시스템의 각 공정별 에너지 인텐시티를 분석하였으며, 시스템 다이나믹스를 적용하여 다양한 도시 여건(인구, lpcd, 누수율, 취수원, 에너지 인텐시티)에서 외부영향(기후변화, 도시화)과 운영효율 변동(운영효율 향상, 신시술 도입)에 따른 도시 물순환 시스템 내 자원 사용 및 이동을 분석하였다. 에너지 인텐시티는 전체 도시 물순환 시스템, 상수 시스템, 하수시스템에서 각각 2.334 kWh/m³, 1.029 kWh/m³, 1.024 kWh/m³를 나타내었으며, 용수사용, 담수화, 재이용 과정에서는 매우 높은 값이 나타났다. 에너지 인틴시티의 값은 외부 영향에 크게 좌우되는 것으로 분석되었으며, 운영효율의 변동에 따라서 물 및 에너지 사용량은 변화하였지만 에너지 인텐시티의 변동은 크지 않았다. 이에 따라 도시 물순환 시스템을 넥서스 관점에서 관리하기 위해서는 에너지 인텐시티 이외에 물 및 에너지 사용량, 유수수량 관점 에너지 인텐시티, 사용수량 관점 에너지 인텐시티를 종합적으로 고려하는 것이 필요하다.

• Journal of Korean Society of Forest Science
• /
• v.112 no.2
• /
• pp.248-257
• /
• 2023

Because of various societal changes, such as the recent improvement in income levels and extension of the flexible work system, the demand for forest recreation activities and their use patterns are undergoing a change. Accordingly, it is necessary to identify the characteristics of each type through the segmentation of the overall forest recreation and welfare markets and to plan differentiated policies for each market type. This study classifies the forest recreation and welfare activities according to four types of users (i.e., passive usage type, ordinary type, active lover type, and indifferent type) using the Latent Class Analysis and examines their demographic and socioeconomic characteristics to explain the differences between the groups. Three policy implications were derived from the results obtained: 1) the group experiencing forest recreation welfare is subdivided; 2) the socioeconomic characteristics that distinguish the groups undertaking forest recreation activities were identified; and 3) the policy targets and characteristics that can increase the experience of forest recreation welfare were identified. This study is insightful as it suggests differentiated policies for each group and proposes policy measures to move to the desirable group.

• Design Convergence Study
• /
• v.15 no.2
• /
• pp.1-15
• /
• 2016

The purpose of this study is to identify the characteristics of aging-friendly technology that can be integrated into future smart homes in an aging society. The literature survey and content analysis method were used to collect and analyze data. Papers of the international journal ICOST (International Conference on Smart Homes and Health Telematics) that professionally deal with converged technologies were analysis units. Sixty-five papers among 215 papers published from 2007 through 2014 were selected on the basis of end-users orientation. Totally, out of 65 papers 76 technology items were extracted. Characteristics of those technologies were analyzed focusing on purpose and application methods. As results, in terms of purpose, the technologies were oriented to provide both of psychological and physiological support to the users, focusing on the safety, convenience, health and entertainment to extend independent life of the elderly. Among the application method such as building structure, furniture, product, wearable device and free movable robot, product were dominant. Through those results, the aging-friendly technology is expected to alleviate a wide range of issues in aging society.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.18-18
• /
• 2020

금강하구와 같이 막힌 대형 하구는 근본적으로 복잡한 물리화학적 프로세스와 다양한 이해당사자간의 갈등 때문에 단편적인 접근으로는 관리의 한계가 있을 수밖에 없는 환경이다. 더욱이 하굿둑의 갑문을 열어 하구순환을 복원하는 해수유통 사업은 그 자체가 가지는 환경적 효용성의 확인과 기존 용수이용 체계의 전면적 변화와 관련된 상당한 비용이 요구된다는 면에서 특별한 논란이 되어왔다. 본 연구는 해수유통으로 인한 금강 하구해역의 수리, 수질, 퇴적 및 생태 등 다양한 환경적 변화를 집중적인 현장 조사 및 통합 모델링을 통해 정량적으로 예측하고 이를 지역 이해당사자 논의 체계인 금강하구해역정책협의회에 제공함으로써 해수유통과 연관된 이해당사자 간 갈등을 완화하고 과학적 자료에 근거한 정책결정을 지원하기 위해 수행되었다(과학-정책의 통합). 조사연구는 하굿둑 갑문 개방에 따른 다양한 시나리오별로 해수유통으로 인한 영향을 하구해역(하굿둑의 바다 쪽 해역)과 하구호(하굿둑의 하천 쪽 담수역)에 미치는 영향을 검토하였다(하천-하구호-연안해역의 공간적 통합). 시나리오는 하구해역정책협의회의 요구사항을 반영하여 개발되었고, 시나리오별 영향파악은 수리(조석 및 파랑 포함), 퇴적물 이동 및 수질은 Delft3D모델을, 유역으로부터의 유량 및 오염물질 부하량은 분포형 비점모델인 STREAM을, 김양식과 이매패류 생산량을 추정하는 생태모델은 통계기반의 포인트 모델을 통해 이루어졌다(수리-퇴적-수질-생태계 모델의 통합). 시나리오 분석결과, 상시 해수유통을 제외한 대부분의 하굿둑 부분 개방 시나리오에서는 하구해역에 미치는 영향이 제한적으로 나타나는 것으로 예측되었다. 이는 하굿둑 부분 개방에 따른 담수방류량 및 이에 대응하는 해수의 유동패턴이 현재와 크게 달라지지 않음을 반영한다. 반면 하구호에 대한 영향은 하구해역에 비해 하굿둑 갑문 운영시나리오에 따라 상대적으로 민감하게 변화되는 것으로 예측되었다. 갑문의 부분 개방을 통해 하구호의 수위를 현재 대비 5cm~50cm 변화시켰을 때 1psu 염분의 저층에서의 침투거리는 하굿둑 상류 6km~18km 정도로 예측되었고 갑문 개방에 의해 염분 침투 거리의 조절이 가능할 것으로 나타났다. 단, 사용된 통합모델은 수년간의 검보정 과정을 반복적으로 수행하여 예측의 정확도를 지속적으로 개선하였지만, 모델운영 자체의 불확실성을 고려할 때 실제 해수유통을 위해서는 시범개방을 통해 모델 예측의 결과를 확인할 필요가 있다.

• The Korean Journal of Franchise Management
• /
• v.2 no.1
• /
• pp.1-45
• /
• 2011

Franchising is one of the most effective strategies in promoting business establishment, creating jobs and accelerating economic growth. Thus, many countries in the world foster franchise business and support the franchise industry by enacting relevant regulations and making development policies. Vietnam deregulated franchise market through making franchise-related laws on June, 2005 and after joining WTO in 2009, Vietnamese government opened up the franchise market by letting foreign investors acquire 100% share on their investment in Vietnam. In addition, Vietnamese consumer's rapidly growing income and increasing liking for western life style make Vietnamese franchise market attractive and profitable. For this reason, many global franchise firms are competing to get the best seats. However, there has been no Korean attempt to profoundly study over Vietnamese franchise market. The purpose of this study would focus on analyzing and evaluating cases of Korean and international firms which successfully entered Vietnamese market and on proposing effective strategies and key success factors. The result of this study indicates that the companies successfully settled down to the Vietnamese market have developed marketing strategies which fit their strengths, and their common key success factors are 1) making the best of first mover advantage, 2) adapting themselves to local tastes and behaviors via systematic and scientific market research, 3) making thorough supervision on safety and hygiene, 4) doing strategy of firstly directly being operated and settled down by franchisor, later being managed by franchisees, 5) benchmarking both successful and failed cases in other countries.

• Korean Journal of School Psychology
• /
• v.15 no.2
• /
• pp.243-264
• /
• 2018

The study examined the influence of socially-prescribed perfectionism (SPP) on depression and social anxiety, and further investigated the mediating effects of self-focused attention and self-criticism. The questionnaires designed to measure multidimensional perfectionism, social anxiety, depression, self-focused attention, self-criticism scale for adolescents were administered twice at an interval of three weeks to 273 students (83 men, 190 women) enrolled at high schools in Gyeonggi-do Province. The findings for the present study were as follows. First, SPP, depression, social anxiety, self-focused attention, and self-criticism showed all positive correlations. Second, the mediation effect from the SPP to depression via self-focused attention was statistically significant, whereas the indirect effect from the SPP to depression via self-criticism was not. Third, the pattern in depression was the same in social anxiety. The results provide indirect support for the social anxiety cognitive model (Clark & Wells) with regards to social anxiety particularly in Korean high school students. Finally, the implications and limitations of this study and suggestions for future research were discussed.

• Information Systems Review
• /
• v.21 no.2
• /
• pp.1-27
• /
• 2019

It is possible to provide Smart Tourism Service through the development of information technology. It is necessary for the tourism industry to understand and utilize Big Data that has tourists' consumption patterns and service usage patterns in order to continuously create a new business model by converging with other industries. This study suggests to activate Jeju Smart Tourism by analyzing Big Data based on credit card usage records and location of tourists in Jeju. The results of the study show that First, the percentage of Chinese tourists visiting Jeju has decreased because of the effect of THAAD. Second, Consumption pattern of Chinese tourists is mostly occurring in the northern areas where airports and duty-free shops are located, while one in other regions is very low. The regional economy of Jeju City and Seogwipo City shows a overall stagnation, without changes in policy, existing consumption trends and growth rates will continue in line with regional characteristics. Third, we need a policy that young people flow into by building Jeju Multi-complex Mall where they can eat, drink, and go shopping at once because the number of young tourists and the price they spend are increasing. Furthermore, it is necessary to provide services for life-support related to weather, shopping, traffic, and facilities etc. through analyzing Wi-Fi usage location. Based on the results, we suggests the marketing strategies and public policies for understanding Jeju tourists' patterns and stimulating Jeju tourism industry.