• Journal of the Korean Society for information Management
• /
• v.32 no.3
• /
• pp.377-396
• /
• 2015

This study proposed a key-frame automatic extraction method for video storyboard surrogates based on users' cognitive responses, EEG signals and discriminant analysis. Using twenty participants, we examined which ERP pattern is suitable for each step, assuming that there are five image recognition and process steps (stimuli attention, stimuli perception, memory retrieval, stimuli/memory comparison, relevance judgement). As a result, we found that each step has a suitable ERP pattern, such as N100, P200, N400, P3b, and P600. Moreover, we also found that the peak amplitude of left parietal lobe (P7) and the latency of FP2 are important variables in distinguishing among relevant, partial, and non-relevant frames. Using these variables, we conducted a discriminant analysis to classify between relevant and non-relevant frames.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.11 no.3
• /
• pp.1-12
• /
• 2008

This study was to develop the analyzing system for watershed characteristics data. The developed system has laid the foundation for preparing the input data in various hydrological models. The algorithms and items included in the system were derived from benchmarking of the GIS WEASEL, which used to prepare the input data of PRMS (Precipitation-Runoff Modeling System). The system was designed to develop and maintain the organized system using both object-oriented UML (Unified Modeling Language) and GIS component-based ArcObjects.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.3
• /
• pp.1086-1103
• /
• 2020

The automatic identification and classification of image-based weld defects is a difficult task due to the complex texture of the X-ray images of the weld defect. Several depth learning methods for automatically identifying welds were proposed and tested. In this work, four different depth convolutional neural networks were evaluated and compared on the 1631 image set. The concavity, undercut, bar defects, circular defects, unfused defects and incomplete penetration in the weld image 6 different types of defects are classified. Another contribution of this paper is to train a CNN model "RayNet" for the dataset from scratch. In the experiment part, the parameters of convolution operation are compared and analyzed, in which the experimental part performs a comparative analysis of various parameters in the convolution operation, compares the size of the input image, gives the classification results for each defect, and finally shows the partial feature map during feature extraction with the classification accuracy reaching 96.5%, which is 6.6% higher than the classification accuracy of other existing fine-tuned models, and even improves the classification accuracy compared with the traditional image processing methods, and also proves that the model trained from scratch also has a good performance on small-scale data sets. Our proposed method can assist the evaluators in classifying pipeline welding defects.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.1
• /
• pp.1-7
• /
• 2011

Terrain, building location and area, and building shape information is in need of implementing 3D map. This paper proposes a method of extracting a building area by an improved semi-automatic snake algorithm. The method consists of 3-stage: pre-processing, initializing control points, and applying an improved snake algorithm. In the first stage, after transforming a satellite image to a gray image and detecting the approximate edge of the gray image, the method combines the gray image and the edge. In the second stage, the user looks for the center point of a building and the system sets the circular or rectangular initial control points by an procedural method. In the third stage, the enhanced snake algorithm extracts the building area. In particular, this paper sets the one tenn of the snake in a new way in order to use the proposed method for specializing building area extraction. Finally, this paper evaluated the performance of the proposed method using sky view satellite image and it showed that the matching percentage to the exact building area is 75%.

• Journal of Internet Computing and Services
• /
• v.5 no.1
• /
• pp.1-13
• /
• 2004

The effect of Application system with class units is not sufficient because of independency and reuse of Component elements due to component abstraction based on only source code. Therefore We need to apply design pattern approach to represent not only the problem abstraction but also information and relationship between system elements for generic solutions of specific domain, Also, it is essential to software reverse engineering acquiring the correct understandings of the system through examining the existing systems and utilizing the acquired knowledges as reusable resources. In this paper, the extraction algorithm with JAVA and the validity of applying reverse engineering with extracting design patterns from source codes of the existing object-oriented system; are devised. The architecture of automatic tool is designed and implemented for 1) automatic extraction of design patterns and 2) reuse tool for retrieving, editing and rebuilding of design patterns.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.2 s.308
• /
• pp.63-71
• /
• 2006

Denoising and reconstruction of color images are extensively studied in the field of computer vision and image processing. Especially, denoising and reconstruction of color face images are more difficult than those of natural images because of the structural characteristics of human faces as well as the subtleties of color interactions. In this paper, we propose a denoising method based on PCA reconstruction for removing complex color noise on human faces, which is not easy to remove by using vectorial color filters. The proposed method is composed of the following five steps: training of canonical eigenface space using PCA, automatic extraction of facial features using active appearance model, relishing of reconstructed color image using bilateral filter, extraction of noise regions using the variance of training data, and reconstruction using partial information of input images (except the noise regions) and blending of the reconstructed image with the original image. Experimental results show that the proposed denoising method maintains the structural characteristics of input faces, while efficiently removing complex color noise.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.4
• /
• pp.374-380
• /
• 2019

Containers are fitted with various placards on the surface to indicate the risk of cargo. If the containers are loaded with dangerous goods, care should be taken in handling the containers. Therefore, as part of the port automation system, there is a demand for automatic placard recognition. In this paper, proposed is a method to extract placard areas from a container image, which is the first part of the placard recognition system. The fact that placards are of various types but all have a diamond shape can be an advantage in recognition. However, it is a disadvantage in recognition that the placards can be distorted in various ways because the container surface is not flat. When the proposed method was applied to actual images, type I error did not occur. In addition, since the shape feature of the object and basic image operations are used to extract regions of interest, it can be applied to various shape-based region extraction problems.

• Journal of the Korean Society for Library and Information Science
• /
• v.55 no.3
• /
• pp.343-372
• /
• 2021

The purpose of this study is to select keywords that can recommend books by personality type, and to test whether the chosen keywords can be actually used in the categorization and customized recommendation of books for each personality type. To achieve the research goal, this study chose books that match the level of fifth and sixth grade elementary school students and first grade middle school students and commissioned an expert group to categorize the books into groups that are preferred by each personality type. As a result of the classification, half of the books in which more than five experts agreed showed high consensus. In addition, the results of classifying books by personality type with keywords extracted by the automatic word extraction system by collecting the book review data of the selected books were similar to the results of the final judgement by the expert group, except for a few books. In conclusion, this study proved that it is possible to classify and recommend books that are likely to be preferred by different personality types using review keywords.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.4
• /
• pp.275-287
• /
• 2024

Research on the integration of Knowledge Graphs (KGs) and Language Models (LMs) has been consistently explored over the years. However, studies focusing on the automatic generation of text using the structured knowledge from KGs have not been as widely developed. In this study, we propose a methodology for automatically generating specific domain-related research items (Related Work) at a level comparable to existing papers. This methodology involves: 1) selecting optimal prompts, 2) extracting triples through a four-step refinement process, 3) constructing a knowledge graph, and 4) automatically generating related research. The proposed approach utilizes GPT-4, one of the large language models (LLMs), and is desigend to automatically generate related research by applying the four-step refinement process. The model demonstrated performance metrics of 17.3, 14.1, and 4.2 in Triple extraction across #Supp, #Cont, and Fluency, respectively. According to the GPT-4 automatic evaluation criteria, the model's performamce improved from 88.5 points vefore refinement to 96.5 points agter refinement out of 100, indicating a significant capability to automatically generate related research at a level similar to that of existing papers.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06a
• /
• pp.188-190
• /
• 2012

인용 레코드 추출은 인용 색인 구축의 모든 프로세스 입력으로 사용되기 때문에 이후의 과정에 미칠 수 있는 부작용을 고려해서 최대한 정확한 정보가 추출되어야 한다. 본 논문에서는 수집한 논문의 참고문헌 영역을 인식하고 이를 참고문헌 영역 내의 특징들을 이용하여 인용 레코드를 추출하는 템플릿 기반 인용 레코드 추출을 제안한다. 제안된 추출 방법은 기존 방법보다 18% 성능이 증가했으며 전체 인용 레코드에 대한 추출성능은 0.98(F1)의 성능을 보인다. 이는 향후 논문 저장소에서 논문을 대상으로 영역 인식을 통한 정보 추출에 유용하게 활용될 수 있으리라 기대된다.