• The Transactions of the Korea Information Processing Society
• /
• v.13 no.4
• /
• pp.181-188
• /
• 2024

Traditional profanity detection methods have limitations in identifying intentionally altered profanities. This paper introduces a new method based on Named Entity Recognition, a subfield of Natural Language Processing. We developed a profanity detection technique using sequence labeling, for which we constructed a dataset by labeling some profanities in Korean malicious comments and conducted experiments. Additionally, to enhance the model's performance, we augmented the dataset by labeling parts of a Korean hate speech dataset using one of the large language models, ChatGPT, and conducted training. During this process, we confirmed that filtering the dataset created by the large language model by humans alone could improve performance. This suggests that human oversight is still necessary in the dataset augmentation process.

• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.818-823
• /
• 2006

본 연구에서는 디지털 콘텐츠와 물리적 미디어가 유기적으로 결합할 수 있는 가능성을 모색하고자, 입출력 일체화 개체인 'I/O 셀(Cell)'과 그 집합체인 '텐저블 게임 콘솔(Tangible Game Console)'을 개발하였다. I/O 셀은 여러 종류의 센서와 액츄에이터가 하나의 물리적 노드 내에 통합되어 있는 개체로, 이 I/O 셀 다수를 집적하면 다양한 크기와 용도의 엔터테인먼트 미디어를 구성하는 것이 가능하다. 그 중 하나로 개발된 텐저블 게임 콘솔은 일종의 센서-액츄에이터 집합체로 셀들 사이의 연계를 통해 다양한 물리적 정보를 인식하고 표현할 수 있다. 텐저블 게임 콘솔은 기존 게임 콘텐츠의 주요한 표현 수단이었던 이미지와 텍스트 정보 대신에 물리적 정보를 통해 게임 콘텐츠를 구성하며, 조작대상과 조작수단이 일체화되어 있고, 또한 음악, 퍼즐, 액션 등 다양한 장르의 게임 콘텐츠가 적용 가능하다는 특징을 가지고 있다. 텐저블 게임 콘솔에서는 불, 바람, 물, 생명체 등의 물리적 현상을 형상/움직임/소리/빛 등을 통해 재현하며, 사용자는 이를 때리기/바람불기/흔들기/기울이기 등의 직접적인 조작을 통해 체험할 수 있다. 또한 물리적인 상호작용 과정이 일정한 게임 규칙을 가지고 구성되어 있어 지속성 있는 놀이 경험을 제공한다. 본 연구에서는 텐저블 게임 콘솔의 콘텐츠로 음악, 액션, 대전게임 등을 개발하였으며, 이 콘텐츠들의 놀이경험을 평가함으로써, 텐저블 게임 콘솔의 발전 가능성을 확인하였다.

• Journal of Navigation and Port Research
• /
• v.29 no.10 s.106
• /
• pp.883-890
• /
• 2005

Database Integration problems has been recognized as a critical issue for effective logistics service in logistics environment. However, researches related to effective methodology for this have been studied theoretically in the DB schema integration, are insufficient in the side of the system realization. The aim of this paper is to present a schema integration technique to integrate DB using XML( eXtensible Markup Language) in the part of practical DB integration, a quantitative methodology for the identification of conflict that is a representative problem on database integration. To achieve this aim, we extracted the entity name and attribute name from DB schema and suggested a quantitative methodology to easily fine name conflict that frequently give raise to a trouble when schema integration, based on the level of semantic similarity between attributes and entities.

• The Journal of Society for e-Business Studies
• /
• v.17 no.3
• /
• pp.1-14
• /
• 2012

As the techniques of electronic and information are advancing, the computing power of a smart phone is becoming more powerful and the storage capacity of a smart phone is becoming larger. As the result, various new useful services are becoming available on smart phones. The context-aware service and mobile augmented reality have recently been the most popular research topics. For those newly developed services, identifying the object in the picture taken by the camera on the phone performs an extremely important role. So, many researches of identifying pictures have been published and most of them are based on the time consuming image recognition techniques. On the contrary, this paper introduces a very fast and effective method of identifying the objects on the photo making use of the sensor data obtained from the smart phone and electronic maps. Our method estimates the line of sight of the camera with the location and orientation information provided by the smart phone. Then it finds any element of the map which intersects the line of sight. By investigating those intersecting elements, our method identifies the objects on the photo.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2005.10a
• /
• pp.353-360
• /
• 2005

Database Integration problems has been recognized as a critical issue for effective logistics service in logistics environment. However, researches related to effective methodology for this have been studied theoretically in the DB schema integration, are insufficient in the side of the system realization. The aim of this paper is to present a schema integration technique to integrate DB using XML(eXtensible Markup Language) in the part of practical DB integration, a quantitative methodology for the identification of conflict that is a representative problem on database integration. To achieve this aim, we extracted the entity name and attribute name from DB schema and suggested a quantitative methodology to easily fine name conflict that frequently give raise to a trouble when schema integration, based on the level of semantic similarity between attributes and entities.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.2
• /
• pp.229-235
• /
• 2021

Blockchain, recognized as one of the core technologies of the 4th industrial revolution, is an Internet-based distributed data management system which does not require centralized control. Blockchain is characterized by the integrity and reliability of information, and blockchains can be used where such characteristics are required. Typical applications of blockchain include finance, transaction, and evaluation management. In this paper, we designed a blockchain-based evaluation management system that allows users to freely create and manage evaluation instances. Evaluation managers can create an evaluation instance according to their purpose and allocate evaluation items and evaluators. When the evaluators finish evaluating the evaluation items, the evaluation manager can aggregate the evaluation results for the instance. Someone, who want to perform evaluation for various purposes but do not have an evaluation system, can implement the evaluation system relatively simply by using this system. In addition, due to the characteristic of the blockchain, evaluators cannot modify the evaluation scores they have already recorded, and neither the system administrator nor the evaluation administrator can modify the evaluation scores. For this reason, the reliability of the evaluation increases.

• Journal of Korean Society for Geospatial Information Science
• /
• v.24 no.4
• /
• pp.67-74
• /
• 2016

Drone images with high spatial resolution are emerging as an alternative to previous studies with extraction limits in high density forests. Individual tree in the dense forests were extracted from drone images. To detect the individual tree extracted through the image segmentation process, the image segmentation results were compared between the combination of DSM and all R,G,B band and the combination of DSM and R,G,B band separately. The changes in the tree density of a deciduous forest was experimented by time and image. Especially the image of May when the forests are dense, among the images of March, April, May, the individual tree extraction rate based on the trees surveyed on the site was 50%. The analysis results of the width of crown showed that the RMSE was less than 1.5m, which was the best result. For extraction of the experimental area, the two sizes of medium and small trees were extracted, and the extraction accuracy of the small trees was higher. The forest tree volume and forest biomass could be estimated if the tree height is extracted based on the above data and the DBH(diameter at breast height) is estimated using the relational expression between crown width and DBH.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.11 no.3
• /
• pp.136-150
• /
• 2008

Many studies have been carried out for automatic extraction of building by LiDAR data or airphoto. Combining the benefits of 3D location information data and shape information data of image can improve the accuracy. So, in this research building recognition algorithm based on contour was used to improve accuracy of building recognition by LiDAR data and elaborate building boundary recognition by airphoto. Building recognition algorithm based on contour can generate building boundary and roof structure information. Also it shows better accuracy of building detection than the existing recognition methods based on TIN or NDSM. Out of creating buffers in regular size on the building boundary which is presumed by contour, this research limits the boundary area of airphoto and elaborate building boundary to fit into edge of airphoto by double active contour. From the result of this research, 3D building boundary will be able to be detected by optimal matching on the constant range of extracted boundary in the future.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.1
• /
• pp.13-20
• /
• 2016

In this paper, we propose an object recognition system that can effectively find out its category, its instance name, and several attributes from the color and depth images of an object with hierarchical feature learning. In the preprocessing stage, our system transforms the depth images of the object into the surface normal vectors, which can represent the shape information of the object more precisely. In the feature learning stage, it extracts a set of patch features and image features from a pair of the color image and the surface normal vector through two-layered learning. And then the system trains a set of independent classification models with a set of labeled feature vectors and the SVM learning algorithm. Through experiments with UW RGB-D Object Dataset, we verify the performance of the proposed object recognition system.

• Journal of KIISE
• /
• v.43 no.7
• /
• pp.773-780
• /
• 2016

Semantic information and features are very important for Semantic Role Labeling(SRL) though many SRL systems based on machine learning mainly adopt lexical and syntactic features. Previous SRL research based on semantic information is very few because using semantic information is very restricted. We proposed the SRL system which adopts semantic information, such as named entity, word sense disambiguation, filtering adjunct role based on sense, synonym cluster, frame extension based on synonym dictionary and joint rule of syntactic-semantic information, and modified verb-specific numbered roles, etc. According to our experimentations, the proposed present method outperforms those of lexical-syntactic based research works by about 3.77 (Korean Propbank) to 8.05 (Exobrain Corpus) F1-scores.