• Proceedings of the KAIS Fall Conference
• /
• 2008.11a
• /
• pp.257-259
• /
• 2008

자연이미지에 내포된 텍스트는 많은 정보를 제공함으로 이를 효과적으로 검출하여 다양한 응용분야에 활용될 수 있다. 본 논문에서는 텍스트 영역의 에지 특징과 국부 최소/최대 변환을 이용하여 자연이미지로부터 텍스트 영역 검출 방법을 제안한다. 에지 검출은 캐니-에지 검출기로 추출하고, 국부 최소/최대 변환을 이용하여 텍스트 영역의 연결성분을 추출한다. 각각 추출된 에지 및 연결성분으로부터 텍스트 영역 후보를 검출하고, 각각의 결과를 결합하여 최종적인 텍스트 후보 영역을 검출하고, 후보 텍스트 영역에 대한 검증을 수행함으로서 최종적인 텍스트 영역을 검출한다. 제안한 방법은 다양한 종류의 자연이미지를 대상으로 실험한 결과, 에지 및 연결성분의 두 가지 특징을 결합함으로서 자연이미지에 존재하는 다양한 형태의 텍스트 영역을 효과적으로 검출하였다.

• Proceedings of the KAIS Fall Conference
• /
• 2010.05a
• /
• pp.191-194
• /
• 2010

최근 모바일 기기 응용 분야에 관한 연구가 활발히 진행되고 있으며 모바일기기로 촬영된 영상에서 텍스트 정보를 추출하고자 하는 많은 연구도 진행되고 있다. 자연영상으로부터 텍스트 인식을 위한 전 단계로 텍스트 영역 검출은 필수적이므로 본 연구는 텍스트 영역의 지역적 연결요소 성분 특징을 고려하여 저 대비 영역에 존재하는 텍스트 영역을 검출하는 방법을 제안한다. 연결요소 성분은 전역 및 지역적 표준편차를 이용하여 연결성분을 추출하고 추출된 연결성분으로부터 후보 텍스트 영역을 검출하고, 후보 텍스트 영역에 대한 검증을 수행함으로서 최종적인 텍스트 영역을 검출한다. 제안한 방법은 낮은 대비를 갖는 자연영상에서 텍스트 영역 검출 성능을 개선하는 결과를 얻었다.

• Proceedings of the KAIS Fall Conference
• /
• 2010.05a
• /
• pp.187-190
• /
• 2010

최근 모바일 기기 응용 분야에 관한 연구가 활발히 진행되고 있으며 모바일기기로 촬영된 영상에서 텍스트 정보를 추출하고자 하는 많은 연구도 진행되고 있다. 자연영상으로부터 텍스트 인식을 위한 전 단계로 텍스트 영역 검출은 필수적이므로 본 연구는 텍스트 영역의 에지-기반 특징을 고려하여 저 대비 영역에 존재하는 텍스트 영역을 검출하는 방법을 제안한다. 에지 성분은 캐니-연산자를 이용하여 검출하고 검출된 에지성분으로부터 후보 텍스트 영역을 검출하고, 후보 텍스트 영역에 대한 검증을 수행함으로서 최종적인 텍스트 영역을 검출한다. 제안한 방법은 낮은 대비를 갖는 자연영상에서 텍스트 영역 검출 성능을 개선하는 결과를 얻었다.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.2
• /
• pp.231-235
• /
• 2008

Recently we witnessed a few scandals on plagiarism among academic paper and novels. Plagiarism on documents is getting worse more frequently. Although plagiarism on English had been studied so long time, we hardly find the systematic and complete studies on plagiarisms in Korean documents. Since the linguistic features of Korean are quite different from those of English, we cannot apply the English-based method to Korean documents directly. In this paper, we propose a new plagiarism detecting method for Korean, and we throughly tested our algorithm with one benchmark Korean text corpus. The proposed method is based on "k-mer" and "local alignment" which locates the region of plagiarized document pairs fast and accurately. Using a Korean corpus which contains more than 10 million words, we establish a probability model (or local alignment score (random similarity by chance). The experiment has shown that our system was quite successful to detect the plagiarized documents.

• The KIPS Transactions:PartB
• /
• v.9B no.4
• /
• pp.509-516
• /
• 2002

This paper proposes an efficient algorithm which recognizes the mixed document image consisting of the images, texts, tables, and straight lines. This system is composed of three steps. The first step is the detection of rotation angle for complementing skewed images, the second is detection of erasing an unnecessary background region and last is the classification of each component included in document images. This algorithm performs preprocessing of detecting rotation angles and correcting documents based on the detected rotation angles in order to minimize the error rate by skewness of the documentation. We detected the rotation angie using only horizontal and vertical components in document images and minimized calculation time by erasing unnecessary background region in the detecting process of component of document. In the next step, we classify various components such as image, text, table and line area included in document images. we applied this method to various document images in order to evaluate the performance of document recognition system and show the successful experimental results.

• Journal of the Korean Society of Safety
• /
• v.31 no.6
• /
• pp.121-128
• /
• 2016

Social Media transformed the mass media based information traffic, and it has become a key resource for finding value in enterprises and public institutions. Particularly, in regards to disaster management, the necessity for public participation policy development through the use of social media is emphasized. National Disaster Management Research Institute developed the Social Big Board, which is a system that monitors social Big Data in real time for purposes of implementing social media disaster management. Social Big Board collects a daily average of 36 million tweets in Korean in real time and automatically filters disaster safety related tweets. The filtered tweets are then automatically categorized into 71 disaster safety types. This real time tweet monitoring system provides various information and insights based on the tweets, such as disaster issues, tweet frequency by region, original tweets, etc. The purpose of using this system is to take advantage of the potential benefits of social media in relations to disaster management. It is a first step towards disaster management that communicates with the people that allows us to hear the voice of the people concerning disaster issues and also understand their emotions at the same time. In this paper, Korean language text mining based Social Big Board will be briefly introduced, and disaster issue detection model, which is key algorithms, will be described. Disaster issues are divided into two categories: potential issues, which refers to abnormal signs prior to disaster events, and occurrence issues, which is a notification of disaster events. The detection models of these two categories are defined and the performance of the models are compared and evaluated.

• The KIPS Transactions:PartB
• /
• v.13B no.6 s.109
• /
• pp.591-600
• /
• 2006

This paper describes the development and implementation of a algorithm to decompose word images from image regions mixed text/graphics in document images using statistical analyses. To decompose word images from image regions, the character components need to be separated from graphic components. For this process, we propose a method to separate them with an analysis of box-plot using a statistics of structural components. An accuracy of this method is not sensitive to the changes of images because the criterion of separation is defined by the statistics of components. And then the character regions are determined by analyzing a local crowdedness of the separated character components. finally, we devide the character regions into text lines and word images using projection profile analysis, gap clustering, special symbol detection, etc. The proposed system could reduce the influence resulted from the changes of images because it uses the criterion based on the statistics of image regions. Also, we made an experiment with the proposed method in document image processing system for keyword spotting and showed the necessity of studying for the proposed method.

• Journal of KIISE
• /
• v.43 no.7
• /
• pp.795-800
• /
• 2016

In image classification, recent CNNs compete with human performance. However, there are limitations in more general recognition. Herein we deal with indoor images that contain too much information to be directly processed and require information reduction before recognition. To reduce the amount of data processing, typically variational inference or variational Bayesian methods are suggested for object detection. However, these methods suffer from the difficulty of marginalizing over the given space. In this study, we propose an image-text integrated recognition system using active vision based on Spatial Transformer Networks. The system attempts to efficiently sample a partial region of a given image for a given language information. Our experimental results demonstrate a significant improvement over traditional approaches. We also discuss the results of qualitative analysis of sampled images, model characteristics, and its limitations.

• Journal of the Korea Computer Graphics Society
• /
• v.23 no.3
• /
• pp.65-75
• /
• 2017

As a 360-degree image carries information of all directions, it often has too much information. Moreover, in order to investigate a 360-degree image on a 2D display, a user has to either click and drag the image with a mouse, or project it to a 2D panorama image, which inevitably introduces severe distortions. In consequence, investigating a 360-degree image and finding an object of interest in such a 360-degree image could be a tedious task. To resolve this issue, this paper proposes a method to find a region of interest and produces a 2D naturally looking image from a given 360-degree image that best matches a description given by a user in a natural language sentence. Our method also considers photo composition so that the resulting image is aesthetically pleasing. Our method first converts a 360-degree image to a 2D cubemap. As objects in a 360-degree image may appear distorted or split into multiple pieces in a typical cubemap, leading to failure of detection of such objects, we introduce a modified cubemap. Then our method applies a Long Short Term Memory (LSTM) network based object detection method to find a region of interest with a given natural language sentence. Finally, our method produces an image that contains the detected region, and also has aesthetically pleasing composition.

• Proceedings of the KAIS Fall Conference
• /
• 2009.05a
• /
• pp.40-43
• /
• 2009

최근 모바일 영상기반 응용 분야에 관한 연구가 활발히 진행되고 있으며 모바일기기로 촬영된 영상에서 문자정보를 추출하고자 하는 많은 연구도 진행되고 있다. 자연이미지로부터 문자정보를 추출을 위한 전단계로 문자영역 검출이 필수적이다. 본 연구는 문자영역의 지역적 에지 및 연결요소 특징을 고려하여 조명 및 복잡한 배경에서도 문자영역을 검출하는 방법을 제안한다. 에지 검출은 캐니-에지 검출기로 추출하고, RGB 컬러분포 패턴을 분석하여 컬러 양자화를 함으로서 연결성분을 추출한다. 각각 추출된 에지 및 연결성분으로부터 문자후보 영역을 검출하고, 각각의 결과를 결합하여 최종적인 문자 후보 영역을 검출하고, 문자 후보 영역에 대한 검증을 수행함으로서 최종적인 문자영역을 검출한다. 제안한 방법은 다양한 환경에서 얻어진 자연이미지를 대상으로 실험한 결과, 에지 및 연결성분의 두 가지 특징을 결합함으로서 자연이미지에 존재하는 다양한 형태의 문자영역을 효과적으로 검출하였다.