• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.47-53
• /
• 2020

문서 요약은 주어진 문서에서 핵심 내용만을 남긴 간결한 요약문을 생성하는 일로 자연어처리의 주요 분야 중 하나이다. 최근 방대한 데이터로부터 심층 신경망 표상을 학습하는 기술의 발전으로 문서 요약 기술이 급진적으로 진화했다. 이러한 데이터 기반 접근 방식에는 모델의 학습을 위한 양질의 데이터가 필요하다. 그러나 한국어와 같이 잘 알려지지 않은 언어에 대해서는 데이터의 획득이 쉽지 않고, 이를 구축하는 것은 많은 시간과 비용을 필요로 한다. 본 논문에서는 한국어 문서 요약을 위한 대용량 데이터셋을 소개한다. 데이터셋은 206,822개의 기사-요약 쌍으로 구성되며, 요약은 표제 형식의 여러 문장으로 되어 있다. 우리는 구축한 학습 데이터의 적합성을 검증하기 위해 수동 평가 및 여러 주요 속성에 대해 분석하고, 기존 여러 문서 요약 시스템에 학습 및 평가하여 향후 문서 요약 벤치마크 데이터셋으로써 기준선을 제시한다. 데이터셋은 https://github.com/hong8e/KHS.git의 스크립트를 통해 내려받을 수 있다.

• Korean Journal of Radiology
• /
• v.22 no.12
• /
• pp.2073-2081
• /
• 2021

Deep learning-based applications have great potential to enhance the quality of medical services. The power of deep learning depends on open databases and innovation. Radiologists can act as important mediators between deep learning and medicine by simultaneously playing pioneering and gatekeeping roles. The application of deep learning technology in medicine is sometimes restricted by ethical or legal issues, including patient privacy and confidentiality, data ownership, and limitations in patient agreement. In this paper, we present an open platform, MI2RLNet, for sharing source code and various pre-trained weights for models to use in downstream tasks, including education, application, and transfer learning, to encourage deep learning research in radiology. In addition, we describe how to use this open platform in the GitHub environment. Our source code and models may contribute to further deep learning research in radiology, which may facilitate applications in medicine and healthcare, especially in medical imaging, in the near future. All code is available at https://github.com/mi2rl/MI2RLNet.

• Korean Journal of Radiology
• /
• v.25 no.7
• /
• pp.603-612
• /
• 2024

Artificial intelligence (AI) is rapidly gaining recognition in the radiology domain as a greater number of radiologists are becoming AI-literate. However, the adoption and implementation of AI solutions in clinical settings have been slow, with points of contention. A group of AI users comprising mainly clinical radiologists across various Asian countries, including India, Japan, Malaysia, Singapore, Taiwan, Thailand, and Uzbekistan, formed the working group. This study aimed to draft position statements regarding the application and clinical deployment of AI in radiology. The primary aim is to raise awareness among the general public, promote professional interest and discussion, clarify ethical considerations when implementing AI technology, and engage the radiology profession in the ever-changing clinical practice. These position statements highlight pertinent issues that need to be addressed between care providers and care recipients. More importantly, this will help legalize the use of non-human instruments in clinical deployment without compromising ethical considerations, decision-making precision, and clinical professional standards. We base our study on four main principles of medical care-respect for patient autonomy, beneficence, non-maleficence, and justice.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.17 no.1
• /
• pp.144-158
• /
• 2014

This research presented a conceptual design of damage assessment inventory for efficient response to natural disaster damage of infrastructure close to the river. It consists of classification and categorization of facilities for accomplishing the conceptual design of inventory for damage of infrastructure close to the river. However, there are arising problems of efficient management on disaster, such as poor management of data facilities and constructions which is managed by the different types of government departments. Therefore, this research presented conceptual models of damage assessment inventory on risks of damage infrastructure close to the river using the United states' HAZUS-MH to analyze damage facilities, type of asset classification, classification of domestic facilities and guidelines for computing the value of assets. Conceptual models of inventory this research presented is to be used on the data for damage response on protected inland damage assessment and to increase efficiency for evaluating detailed damage amount of private property by natural disaster and to establish a restoration plan.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.18 no.4
• /
• pp.131-144
• /
• 2015

This research proposes a prototype for visualizing radar rainfall data using the unity 3D engine. The mashup of radar data with topographic information is necessary for the 3D visualization of the radar data with high quality. However, the mashup of a huge amount of radar data and topographic data causes the overload of data processing and low quality of the visualization results. This research utilized the Unitiy 3D engine, a widely used engine in the game industry, for visualizing the 3D topographic data such as the satellite imagery/the DEM(Digital Elevation Model) and radar rainfall data. The satellite image segmentation technique and the image texture layer mashup technique are employed to construct the 3D visualization system prototype based on the topographic information. The developed protype will be applied to the disaster-prevention works by providing the radar rainfall data with the 3D visualization based on the topographic information.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.14 no.1
• /
• pp.118-129
• /
• 2011

This Study has developed the information system of the rivers based on 3D image GIS by converging the latest information technology of GIS(Geographic Information System), RS(Remote Sensing), GNSS(Global Navigation Satellite System), aerial laser survey(LiDAR) with real time network technology in order to understand the current situation of all the four major rivers and support the administrative management system. The said information system acquires the high resolution aerial photographs of 25cm, aerial laser survey and water depth surveying data to express precise space information on the whole Youngsan River which is the leading project site out of the four river sites. Monitoring the site is made available on the transporting means such as a helicopter, boat or a bus in connection with locational coordinate tracking skill for the moving objects in real time using GNSS. It makes monitoring all the information on the four river job sites available at a glance, which can obtain the reliability of the people to such vast areas along with enhancing the recognition of the people by publicity of four Rivers Revitalizing Project and reports thereof.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.14 no.1
• /
• pp.84-93
• /
• 2011

Levee line mapping is critical to the protection of environments in river zones, the prevention of river flood and the development of river zones. Use of the remote sensing data such as LiDAR and aerial orthoimage is efficient for river mapping due to their accessibility and higher accuracy in horizontal and vertical direction. Airborne laser scanning (LiDAR) has been used for river zone mapping due to its ability to penetrate shallow water and its high vertical accuracy. Use of image source is also efficient for extraction of features by analysis of its image source. Therefore, aerial orthoimage also have been used for river zone mapping tasks due to its image source and its higher accuracy in horizontal direction. Due to these advantages, in this paper, research on three dimensional levee line mapping is implemented using LiDAR and aerial orthoimage separately. Accuracy measurement is implemented for both extracted lines generated by each data using the ground truths and statistical comparison is implemented between two measurement results. Statistical results show that the generated 3D levee line using LiDAR data has higher accuracy than the generated 3D levee line using aerial orthoimage in horizontal direction and vertical direction.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.21 no.2
• /
• pp.68-79
• /
• 2018

In general, Top Of Atmosphere(TOA) brightness temperature is closely related to air temperature. Brightness temperature can be derived from the Thermal Infra-Red Sensors (TIRS) of the earth observation satellites such as the Landsat series. The TIRS instrument of the Landsat-8 satellite collects the two spectral bands (Bands 10 and 11) that measure brightness temperature. In this research, the relationship between the air temperature data measured by the weather stations in Seoul, South Korea and the brightness temperature data separately derived from Bands 10 and 11 of the Landsat-8 satellite were assessed in the different seasons through the correlation analysis. The statistical results led to the following conclusions. First, brightness temperature is closely related to air temperature in order of Spring, Autumn, Winter and Summer. Second, when air temperature increases, brightness temperature also increases in Spring, Autumn and Winter but decreases in Summer. Third, Band 10 has a closer relationship to air temperature than Band 11.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.19 no.1
• /
• pp.141-153
• /
• 2016

The increase in frequency and scale of natural disasters is the typical negative examples of the global climate change and the change of the human living environment in cities. The damage caused by natural disasters in particular including human and physical damage is directly linked to the safety and properties of citizens. Besides, the damage may occur to SOC facilities directly or indirectly. The SOC facilities damaged by disasters cause infringement of citizens safety rights. Therefore, a plan to provide prompt and effective risk map information by linking a 3D disaster information display system, which handles the information of the damage that may occur to SOC facilities such as bridges, beams, and dams at the time of disasters, with an on-site assistance application is suggested in this study. The prompt provision of risk map information is defined as a dynamic expression technology in this study. Also, disaster information is processed and compressed with a module developed to be linked organically to a system that displays disaster information, a proliferation system that can use the information on site, and a DB system that constructs a relationship with the information. Based on the module, the effective disaster information compression plan will be prepared, and the prompt information transmission system will be secured in the future.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.19 no.1
• /
• pp.180-196
• /
• 2016

To minimize the damage from disasters, various aspects of prevention, preparation, and response, etc. are being managed. Even though prevention and preparation are well conducted, irresistible calamities such as natural disasters may cause unexpected damage. Therefore, a system that can share the identical disaster information based on prompt disaster management and prediction must be developed and constructed for integrated disaster management. Especially, for a prompt disaster response, the same information needs to be shared between the related organization and the disaster prevention personnel such as on-site officials. Recent disaster management systems use high-capacity geographic information or other various factors for accurate disaster predictions. In case of using a recently constructed or researched 3D GIS, the system may not be used in some cases due to conflicts with hardware, etc. Thus, even though response information is secured using prediction simulation in advance, it is essentially difficult in some cases to share the common information when the system cannot be utilized or the extension of the corresponding data cannot be read. Therefore, this study aims to construct a system for dealing with disasters that shares the same prompt and accurate information in compliance with common data formats. The system is expected to reduce the existing disaster response time and minimize human and physical damage by assisting decision making through prompt responses.