• Journal of the Korea Society of Computer and Information
• /
• v.14 no.12
• /
• pp.89-96
• /
• 2009

A lot of information is contained in an image or a movie rather than in a text, and it is very important thing to extract context from them. In this study, we propose a method to detect an open/closed eye and determine the location of a pupil in an eye image which is extracted from a movie. The image is normalized using transformation into bipolarization with white and black color and horizontalizing, and we measure width and height of an eye. With these information, we can determine the open or closed eye and the location of the pupil. Experiments were done with 52 images of eyes from movies using this method, and we get good results with 98% of correctness in detection of open/closed eyes and 95% in detection of pupil's location.

• Journal of Digital Convergence
• /
• v.20 no.1
• /
• pp.285-291
• /
• 2022

The Purpose of this Study is to explore the new direction for Metaverse marketing and I analyze case of Metaverse marketing focusing on jewelry brand and changes of IT technology for Metaverse Roadmap 2.0. Based on the analyzed marketing strategy, jewelry brands compare and study Metaverse marketing cases to draw implications. As a result of the study, successful Metaverse marketing provides a personalized experience in the virtual space and is accompanied by analysis of the customer journey, and this can be confirmed in the case of global brand. As a future research direction, Through in-depth research on marketing ROI(Return On Investment), I contribute to enhancing the competitiveness of jewelry brand.

• Journal of Digestive Cancer Research
• /
• v.10 no.1
• /
• pp.31-38
• /
• 2022

Ulcerative colitis (UC) exhibits chronic intestinal inflammatory conditions with cycles of relapse and remission. The incidence is rapidly growing in Asian countries including South Korea possibly due to changes in lifestyles. Although the etiology of inflammatory bowel disease is inconclusive, gut microbiota composition is considered a critical factor involved in the pathogenesis of UC. The overgrowth of pathogenic bacteria evokes hyper-immune responses in gut epithelium causing tissue inflammation and damage. Also, failure to regulate gut epithelium integrity due to chronic inflammation and mucus depletion accelerates bacterial translocation aggravating immune dysregulation. Gut microbiota composition responds to the diet in a very rapid manner. Epidemiological studies have indicated that the risk of UC is associated with low plant foods/high animal foods consumption. Several bacterial strains consistently found depleted in UC patients use plant food-originated dietary fiber producing short chain fatty acids to maintain epithelial integrity. These bacteria also use mucus layer mucin to keep gut microbiota diversity. These studies partly explain the association between dietary modification of gut microbiota in UC development. Further human intervention trials are required to allow the use of specific bacterial strains in the management of UC.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.507-511
• /
• 2020

오픈 도메인 대화에서 텍스트에 나타난 태도나 성향과 같은 화자의 주관적인 감정정보를 분석하는 것은 사용자들에게서 풍부한 응답을 이끌어 내고 동시에 제공하는 목적으로 사용될 수 있다. 하지만 한국어 감성분석에서 기존의 대부분의 연구들은 긍정과 부정 두개의 클래스 분류만을 다루고 있고 이는 현실 화자의 감정 정보를 정확하게 분석하기에는 어려움이 있다. 또한 최근에 오픈한 다중클래스로된 한국어 대화 감성분석 데이터셋은 중립 클래스가 전체 데이터셋의 절반을 차지하고 일부 클래스는 사용하기에 매우 적은, 다시 말해 클래스 간의 데이터 불균형 문제가 있어 다루기 굉장히 까다롭다. 이 논문에서 우리는 일곱개의 클래스가 존재하는 한국어 대화에서 세션들을 효율적으로 분류하는 기법들에 대해 논의한다. 우리는 극심한 클래스 불균형에도 불구하고 76.56 micro F1을 기록하였다.

• The Journal of Bigdata
• /
• v.6 no.2
• /
• pp.181-195
• /
• 2021

Interest in the transition to MyData beyond the use of personal data is high worldwide. In order for the ideology of MyData to be realized, various useful MyData services would be developed in practice. Above all, convergence must be premised for the development of MyData. This study proposed strategies for activating MyData convergence services from four perspectives (BLTS: Business, Legal, Technology, and Social). In particular, the purpose w as to revitalize domestic convergence services by comparing European policies and services, which are the most pioneering in the field of MyData, with the domestic situation. It is expected that this study will provide balanced and progressive ideas in the MyData industry and policies.

• Journal of Software Assessment and Valuation
• /
• v.15 no.2
• /
• pp.85-93
• /
• 2019

In an open, heterogeneous environment based on machine-to-machine (M2M) interactions, service selection is a critical issue and the concept of social trust can be applied to service selection so that IoT devices can make the best choice for interaction. In this paper, we propose a method for evaluating the trust level of the service and for estimating the QoS of the composite service using a profile created based on social trust relationship in IoT environment. As the service selection is made through quantitative evaluation, it is expected that the result of a more reliable service combination can be obtained.

• Journal of Software Assessment and Valuation
• /
• v.15 no.1
• /
• pp.123-131
• /
• 2019

The Internet of Things (IoT) environment must be able to meet the needs of users by providing access to various services that can be used to develop diverse user applications. However, QoS issues arise due to the characteristics of the IoT environment, such as numerous heterogeneous devices and potential resource constraints. In this paper, we propose a QoS prediction method that reflects trust between users in SOA based IoT. In order to increase the accuracy of QoS prediction, we analyze the trust and distrust relations between users and identify similarities among users and predict QoS based on them. The centrality is calculated to enhance trust relationships. Experimental results show that QoS prediction can be improved.

• Human Ecology Research
• /
• v.61 no.3
• /
• pp.335-348
• /
• 2023

The purpose of this study was to identify: (a) the latent profiles of the difficulties associated with the schoolto-work transition (decline in confidence, mood swings, family disagreements, the burdens of familial expectations, economic hardship, and a lack of support) made by university students, (b) predictors (gender, age, grade, university location, co-residence with parents on weekdays, monthly household income, and parental educational attainment) of these profiles, and (c) how the profiles were associated with mental wellbeing. The participants of this study were 311 senior or above students (164 males and 147 females) under the age of 29, who were unmarried and preparing for employment. The findings of this study were as follows. First, the latent profile analysis revealed three distinct profiles: the "low overall difficulties" type (25.4%), the "moderate overall difficulties" type (49.9%), and the "high overall difficulties" type (24.7%). Second, the factors that predicted each profile included gender, age, co-residence with parents on weekdays, monthly household income, and parental educational attainment. Third, the "low overall difficulties" type demonstrated the highest level of mental well-being (emotional, social, and psychological well-being). This study was significant for examining the latent profiles of the difficulties associated with the school-to-work transition made by university students preparing for employment, while also exploring their mental well-being. Based on the results of this study, practical implications, limitations, and suggestions for further study were discussed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.1094-1095
• /
• 2023

우리나라는 지리적 여건 상 대외무역에 대한 의존도가 높기 때문에, 해상운송에서의 물동량을 예측하여 항만시설을 개발하는 것이 매우 중요하다. 한편 우리나라 컨테이너 운송의 75%는 부산항을 통해 운송되고 있기 때문에 경기 회복을 위해서는 부산항의 경쟁력 강화가 급선무이다. [1] 물동량은 경제적 수입 뿐만 아니라, 지속가능성을 예측하는 측면에서도 가치가 있다. 본 연구에서는 물동량, 경제지수, 기후정보 등 다양한 입력변수와 LSTM 모델을 이용하여 보다 정확한 부산항 컨테이너 물동량 딥러닝 예측모델을 구현하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.483-486
• /
• 2023

최근 금융 분야는 예측 모델의 복잡성으로 인한 블랙박스 문제와 금융 규제에 대한 관심이 높아지고 있다. 이에 따라 금융 업계는 신뢰성과 투명성을 강조하며, 특히 신용평가 분야에서 설명 가능한 모델 연구가 활발히 진행되고 있다. 또한, 해당 분야에서 소수 클래스에 대해 충분히 학습하지 못하고 다수 클래스에 과적합 될 수 있는 데이터 불균형 문제 역시 강조되고 있다. 이는 제 2종 오류(Type 2 Error)를 최소화해야 하는 상황에서 더욱 부각되며, 대출 상환 능력이 낮은 고객을 최대한 식별해야 하는 개인 신용평가 문제에서 매우 중요한 화두로 떠오르고 있다. 본 논문에서는 어텐션 메커니즘을 활용하여 모델의 설명 가능성을 개선하고, 분석 결과를 해석하는 데 도움이 되고자 한다. 더 나아가, SMOTE, GAN, ADASYN 등 총 다섯 가지 데이터 증강 기법을 실험하여, 이를 앙상블 하였을 때 소수 클래스 레이블에 대한 분류 정확도를 크게 개선할 수 있음을 확인하였다.