• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.329-330
• /
• 2020

본 논문에서는 Azure 클라우드 플랫폼의 가상서버 호스팅을 이용해 데이터 수집 환경을 구축하고, Azure에서 제공하는 자동화된 기계학습(Automated Machine Learning, AutoML)을 기반으로 데이터 분석 방법에 관한 연구를 수행했다. 가상 서버 호스팅 환경에 LAMP(Linux, Apache, MySQL, PHP)를 설치하여 데이터 수집환경을 구축했으며, 수집된 데이터를 Azure AutoML에 적용하여 자동화된 기계학습을 수행했다. Azure AutoML은 소모적이고 반복적인 기계학습 모델 개발을 자동화하는 프로세스로써 기계학습 솔루션 구현하는데 시간과 자원(Resource)를 절약할 수 있다. 특히, AutoML은 수집된 데이터를 분류와 회귀 및 예측하는데 있어서 학습점수(Training Score)를 기반으로 보유한 데이터에 가장 적합한 기계학습 모델의 순위를 제공한다. 이는 데이터 분석에 필요한 기계학습 모델을 개발하는데 있어서 개발 초기 단계부터 코드를 설계하지 않아도 되며, 전체 기계학습 시스템을 개발 및 구현하기 전에 모델의 구성과 시스템을 설계해볼 수 있기 때문에 매우 효율적으로 활용될 수 있다. 본 논문에서는 NPU(Neural Processing Unit) 학습에 필요한 데이터 수집 환경에 관한 연구를 수행했으며, Azure AutoML을 기반으로 데이터 분류와 회귀 등 가장 효율적인 알고리즘 선정에 관한 연구를 수행했다.

• Journal of Plant Biology
• /
• v.37 no.2
• /
• pp.203-221
• /
• 1994

Pollen grains of 19 taxa belonging to 12 genera in 5.families of the Korean gymnosperms were investigated by LM, SEM and TEM, and their morphological characters were described and palynological keys were made for each family. From the taxonomic viewpoint, palynological characters are very available to intergeneric and intrageneric classifications since their distinctive differences in pollen shape, size, condition of surface, sculpture pattern and presence of air bladders and projection except for the intrageneric classification of the genus Pinus since the overlapped characters such as pollen size and sculpture pattern. Pollen grains of Picea koraiensis in the family Pinaceae are distinguished from thc genera Pinus and Abies for the reasons that shape of air bladders are irregular, air bladders are bigger than pollen body, and pollen wall is thin. Compared with the genera Abies and Pinus that have air bladders, Picea koraiensis is closely related to Abies in that pollen body wall become gradually thin from pole of proximal face to air bladder, surface of pollen body is not rough, and pollen size is similar to that of the genus Abies. The presence of ektexine and enexine in proximal face, and the connections of the former to air bladder wall and the latter to distal face were elucidated by the investigation of pollen grains with air bladders through TEM. Endexine of the families of Taxaceae, Taxodiaceae and Cupressaceae was also obse-rved as a pile of laminae. From the observation by SEM, sculpture pattern of air bladder was foveolate with fine pits on surface. Sculpture pattern of air bladder is reticulate under LM, but observed by TEM, it resulted from the transmission of netted structure in air bladder.

• The Korean Journal of Mycology
• /
• v.27 no.5 s.92
• /
• pp.341-344
• /
• 1999

The genus Hypocrea, a member of the Hypocreales, has yellow to pale yellow-ochre, perithecial stromata and angular to cuboid ascospores. The isolate of Hypocrea microrufa collected on decayed tropical wood at Malaysia is compared with H. rufa and H. microsplendens. This isolation of H. microrufa will be the first record from tropical region and the first description with scanning electron microscopy. The ornamentation of ascospores used one of main taxonomic keys for identification.

• Korean journal of applied entomology
• /
• v.55 no.2
• /
• pp.103-107
• /
• 2016

Taxonomy of Korean Paracymus Thomson is presented. The genus and two species [Paracymus aeneus (Germar) and P. zaitzevi Shatrovskiy] are identified for the first time in the Korean Peninsula. Habitus and SEM photographs, distribution map, key and diagnoses of the known species are provided.

• Korean journal of applied entomology
• /
• v.47 no.4
• /
• pp.315-331
• /
• 2008

A total of 12 species in six genera of the poneromorph subfamilies group in Korea was reviewed. All investigated specimens are classified into three subfamilies; Amblyoponinae, Proceratinae and Ponerinae. Totally six genera were known and Cryptopone, Hypoponera, Pachycondyla and Ponera belong to the subfamily Ponerinae and Amblyopone and Proceratium belongs to the subfamily Amblyoponinae and Proceratinae respectively. In this study, keys to the all known genera and species of the subfamilies in Korea are presented. A synonymic list, description of a worker and comments on each species are given with SEM photographs of worker.

• The Korean Journal of Archival Studies
• /
• no.14
• /
• pp.153-190
• /
• 2006

Externally, business environment in public institution has being changed as government business reference model(BRM) appeared and business management systems for transparency of a policy decision process are introduced. After Records Automation System started its operation, dissatisfaction grows because of inadequacy in system function and the problems about authenticity of electronic records. With these backgrounds, National Archives and Records Service had carried out 'Information Strategy Planning for Reform to Records Management System' for 5 months from September, 2005. As result, this project reengineers current records management processes and presents the world-class system model. After Records and Archives Management Act was made, the records management in public institution has propelled the concept that paper records are handled by means of the electric data management. In this reformed model, however, we concentrates on the electric records, which have gradually replaced the paper records and investigate on the management methodology considering attributes of electric records. According to this new paradigm, the electric records management raises a new issue in the records management territory. As the major contents of the models connecting with electric records management were analyzed and their significance and bounds were closely reviewed, the aim of this paper is the understanding of the future bearings of the management system. Before the analysis of the reformed models, issues in new business environments and their records management were reviewed. The government's BRM and Business management system prepared the general basis that can manage government's whole results on the online and classify them according to its function. In this points, the model is innovative. However considering the records management, problems such as division into Records Classification, definitions and capturing methods of records management objects, limitations of Records Automation System and so on was identified. For solving these problems, the reformed models that has a records classification system based on the business classification, extended electronic records filing system, added functions for strengthening electric records management and so on was proposed. As regards dramatically improving the role of records center in public institution, searching for the basic management methodology of the records management object from various agency and introducing the detail design to keep documents' authenticity, this model forms the basis of the electric records management system. In spite of these innovations, however, the proposed system for real electric records management era is still in its beginning. In near feature, when the studies is concentrated upon the progress of qualified classifications, records capturing plans for foreign records structures such like administration information system, the further study of the previous preservation technology, the developed prospective of electric records management system will be very bright.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.6
• /
• pp.1331-1336
• /
• 2023

In text-based fish disease classification using machine learning, there is a problem that the input parameters of the machine learning model are too many, but due to performance problems, the input parameters cannot be arbitrarily reduced. This paper proposes a method of optimizing input parameters specialized for Paralichthys olivaceus disease classification using SHAP analysis techniques to solve this problem,. The proposed method includes data preprocessing of disease information extracted from the halibut disease questionnaire by applying the SHAP analysis technique and evaluating a machine learning model using AutoML. Through this, the performance of the input parameters of AutoML is evaluated and the optimal input parameter combination is derived. In this study, the proposed method is expected to be able to maintain the existing performance while reducing the number of input parameters required, which will contribute to enhancing the efficiency and practicality of text-based Paralichthys olivaceus disease classification.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.31 no.2
• /
• pp.139-160
• /
• 2020

Since the Public Records Management Act was enacted in 2007, the administrative information system has already been included in the electronic records production system, and dataset has been subject to record management as a type of electronic records. With the recent revision of the enforcement decree, dataset records management has been enacted. This study analyzes business units related to administrative information systems of public institutions and examines the current status of retention periods estimation. For this purpose, we collected 36 records classification systems from 49 public institutions among the direct management agencies of the National Archives and disaster management agencies. And we discriminated 824 business units related to administrative information system and divided into large and small groups according to types. We also compared the retention period estimation of records. The problems and improvement plans of this study are expected to be used as basic data in preparing the standard of administrative dataset management in the future.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.6
• /
• pp.1229-1236
• /
• 2022

In this paper, we studied accident-aware smart streetlights to prevent secondary accidents when driving on highways. It used Arduino and sensors to inform drivers of weather conditions, incorporated functions such as LED brightness control according to sunlight and night driving vehicles, and used Raspberry Pi camera OpenCV to learn various traffic accidents, natural disasters, and wildlife.

• Electronics and Telecommunications Trends
• /
• v.22 no.4 s.106
• /
• pp.35-42
• /
• 2007

1990년부터 디지털콘텐츠 제작에 사용되고 있는 모션캡처 기술은 제작비 절감, 콘텐츠의 리얼리티(reality) 향상이라는 장점으로 인해 지속적으로 활용되어 왔다. 최근 디지털콘텐츠가 고품질화 함에 따라 리얼리티의 중요도가 매우 커졌으며, 이로 인해 디지털액터를 리얼하게 움직이기 위한 모션캡처 기술의 활용이 지속적으로 증가하고 있다. 본 고에서는 모션캡처를 데이터 추출 방식에 따라 분류하고, 각각의 장단점에 대해 설명한다. 또한 산업계를 중심으로 개발되고 있는 상용화 기술 동향에 대해 기술하고, 향후 발전 방향을 제시한다.