• The Journal of the Korea Contents Association
• /
• v.21 no.6
• /
• pp.701-715
• /
• 2021

The drastic advance of recent deep learning technologies is heavily dependent on training datasets which are essential to train models by themselves with less human efforts. In comparison with the work to design deep learning models, preparing datasets is a long haul; at the moment, in the domain of vision intelligent, datasets are still being made by handwork requiring a lot of time and efforts, where workers need to directly make labels on each image usually with GUI-based labeling tools. In this paper, we overview the current status of vision datasets focusing on what datasets are being shared and how they are prepared with various labeling tools. Particularly, in order to relieve the repetitive and tiring labeling work, we present an interactive smart image annotating system with which the annotation work can be transformed from the direct human-only manual labeling to a correction-after-checking by means of a support of automatic labeling. In an experiment, we show that automatic labeling can greatly improve the productivity of datasets especially reducing time and efforts to specify regions of objects found in images. Finally, we discuss critical issues that we faced in the experiment to our annotation system and describe future work to raise the productivity of image datasets creation for accelerating AI technology.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.16-19
• /
• 2020

본 논문은 국제표준화인 OHDSI OMOP-CDM 의 확장으로 의료영상 표준기반의 R_CDM 으로 변환하고 그 데이터를 기반으로 다기관 임상연구를 위한 관리시스템에 대해 기술한다. 이를 위해 기존 공통데이터모델과 연계에 중점을 두어 DICOM 태그정보를 기반으로 의료영상 표준 모델의 스키마와 다기관 연구를 위한 Report 정보를 포함하여 모델링하였다. 이를 기반으로 머신러닝 기술개발을 위한 데이터 셋 생성과 관리를 위한 웹 기반 시스템 구조와 기능에 대해서 기술한다. 끝으로 구현된 시스템에서 제공하는 웹 서비스 수행 결과를 보인다.

• Digital Contents
• /
• no.10 s.113
• /
• pp.46-51
• /
• 2002

국내 온라인교육 시장이 당초 급성장 할 것이라는 예측과 달리 답보상태를 벗어나지 못하고 있다. 특히 올 초 노동부의 '인터넷통신훈련시행지침'이 개정되면서 자격 요건이 강화돼 많은 업체들이 어려움을 겪어야 했다. 이러한 가운데 업계에서는 침체 및 조정기에 빠져 있는 국내 온라인교육 시장의 활성화를 위해 e-러닝산업협회를 출범시키는 등 다양한 노력을 기울이고 있다. 출범 4년째를 맞은 국내 온라인교육 시장의 현황과 문제점을 짚어 봤다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.10a
• /
• pp.893-895
• /
• 2016

Data center has a lot of management efforts for the facility, energy, and efficient usage monitoring. Data center power management is important to make the data center have reliable service and cost-effective business. In this paper, IoT based energy measurements monitoring which gives support to energy consumption analysis including indoor, outdoor temperature condition. This converged information for energy analysis gives various aspects of energy consumption effects. With IoT big data, energy machine learning system can give the relation of energy components and measurements, it is the key information of the quick energy analysis in the just one month data trend for the prediction and estimation.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.19-21
• /
• 2021

인공지능 기술을 도입한 의료분야에서 진단 및 예측을 위한 관련 연구가 활발하게 진행되고 있다. 특히, 인공지능 기술 적용에 가장 많이 활용되고 있는 의료영상기반 질환 진단 및 예측 연구결과가 다양한 제품으로 출시되고 있다. 의료영상이 활용되는 다양한 질환 중 간 질환은 통증이 적어 조기진단이 어렵다. 본 논문에서는 인공지능을 기반 간 경화증 환자의 판독을 돕기 위한 웹 서비스기반 시스템을 구축하고 진단결과를 보인다. 이를 위해 웹서비스 프로세스를 보이고 각 프로세스의 구동 화면과 최종 결과화면을 보인다. 제안한 서비스를 통해 간 경화증을 조기에 진단하고, 빠른 치료를 통해 환자의 회복에 도움을 줄 수 있을 것으로 기대한다.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2023.05a
• /
• pp.160-161
• /
• 2023

Currently, the ship's passage planning relies on the navigator officer's knowledge and empirical methods. However, as ship autonomous navigation technology has recently developed, automation technology for passage planning has been studied in various ways. In this study, we intend to generate an optimal route for a ship based on Q-learning, one of the reinforcement learning techniques. Reinforcement learning is applied in a way that trains experiences for various situations and makes optimal decisions based on them.

• Journal of Digital Convergence
• /
• v.13 no.7
• /
• pp.9-21
• /
• 2015

Nowadays, within the paradigm shift of e-learning, MOOCs service has been expanded among a number of global leading universities and this has affected our domestic universities' e-learning to look for different possibilities and new challenges. In consideration of our domestic educational condition, it is required to contemplate how university e-learning can be changed and developed with focus on MOOCs from the perspective of convergence development. This study suggests plans for convergence development in university e-learning based on MOOCs with conceptual model. We conducted the studies on relevant literature of university e-learning and MOOCs, expert consultation, SWOT analysis, survey for those involved of e-learning centers, etc. Through this process, we developed a final plan which integrates 'open advanced education course service', 'teaching and learning curation service', 'teaching and learning practice service', 'creative teaching and instruction method development and sharing service', and 'cloud based educational platform support service', etc with the perspective of convergence development. Also we designed convergence development plan based on MOOCs. It is assumed that the result of this research provides advanced plans for development of university e-learning and the base for further discussion of introduction and application of MOOCs service in domestic university.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.5
• /
• pp.231-240
• /
• 2019

The purpose of this study is to understand of flipped learning of vocational high school teachers and to provide basic data of teacher training program development that can support flipped learning classes of vocational high school. We conducted surveys and focus group interviews to solve these research problems. The focus group interview is conducted on four flipped learning teachers who participated in the questionnaire. The results of this study are as follows. First, vocational high school teachers only heard the term 'flipped learning', but they did not have much experience in class. Second, it is necessary to learn contents and pre-learning for basic understanding of flipped learning, method of making learning contents of learning materials that focused on students' motivation and interest in the classroom. Also, we needed a method to induce participation in classroom activities, and a compensation method and evaluation method to maintain it. In addition, they requested specific flipped learning instruction procedures and various learning materials at each stage in accordance with the vocational high school situation. Third, I requested training contents and practice - centered training method that can induce interest and motivation as the training management method for the vocational high school teachers' flipped learning classes. In addition, it required step - by - step training according to the understanding level of flipped learning of high school teacher.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.07a
• /
• pp.273-276
• /
• 2019

본 논문에서는 열 및 전기특성 플라스틱 복합수지와 한글에 특화된 인공지능 기술을 개발하기 위한 조성/물성 정보 복합수지 지식베이스를 구축하고자 국내 특허 문헌에서 화학 용어를 추출하고자 한다. 이를 위해 전문용어가 많이 쓰인 특허 문헌의 특수성을 고려하여 UIMA(Unstructured Information Management Architecture) 규칙 기반의 라이브러리를 사용해 한국어 화학 용어 코퍼스를 구축하고 이를 기반으로 딥러닝 알고리즘 중 하나인 Bidirectional LSTM-CRF를 기반으로 특허 문헌에서 화학 용어를 자동으로 추출하는 알고리즘을 연구하고자 한다.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.112-117
• /
• 2021

기계 학습을 활용하여 요약문을 생성했을 경우, 해당 요약문의 정확도를 측정할 수 있는 도구는 필수적이다. 원문에 대한 요약문의 사실관계 일관성의 파악을 위해 개체명 유사도, 기계 독해를 이용한 질문-답변 생성을 활용한 방법이 시도되었으나, 충분한 데이터 확보가 필요하거나 정확도가 부족하였다. 본 논문은 딥러닝 모델을 기반한 개체명 인식기와 질문-답변쌍 정확도 측정기를 활용하여 생성, 필터링한 질문-답변 쌍에 대해 일치도를 점수화하는 방법을 제안하였다. 이러한 기계적 사실관계 확인 점수와 사람의 평가 점수의 분포를 비교하여 방법의 타당성을 입증하였다.