• 전자공학회논문지CI
• /
• 제47권1호
• /
• pp.73-79
• /
• 2010

감성 지능형 컴퓨팅은 컴퓨터가 학습과 적응을 통하여 인간의 감성을 처리할 수 있는 감성인지 능력을 갖는 것으로 보다 효율적인 인간과 컴퓨터의 상호 작용을 가능하게 한다. 감성 정보들 중 시각과 청각 정보인 음악 이미지는 짧은 시간에 형성되고 기억에 오랫동안 지속되기 때문에 성공적인 마케팅에 있어서 중요한 요인으로 꼽히고 있으며, 인간의 정서를 이해하고 해석하는데 있어서 매우 중요한 역할을 한다. 본 논문에서는 사용자의 감성키워드(짜증, 우울, 차분, 기쁨)를 고려하여 매칭된 음악과 이미지를 검색하는 시스템을 구축하였다. 제안된 시스템은 인간의 감성을 4단계 경우로 상황을 정의하며, 정규화 된 음악과 이미지를 검색하기 위해 음악 이미지 온톨로지와 감성 온톨로지를 사용하였으며, 이미지의 특징정보를 추출, 유사성을 측정하여 원하는 결과를 얻게 하도록 하였다. 또한, 이미지 감성인식정보를 분류하기위해 대응일치분석과 요인분석을 통한 성컬러와 감성어휘를 하나의 공간에 매칭하였다. 실험결과 제안된 시스템은 4가지 감성상태에 대해 82.4%의 매칭율를 가져올 수 있었다.

• 대한임베디드공학회논문지
• /
• 제10권1호
• /
• pp.13-23
• /
• 2015

Recently, a few image-processing based mobile urine testers have actively been studied since the urine-analysis result can be available to the user in real time immediately after the test is done. However, the accuracy of test result can be severely degraded due to variable illumination environments and a variety of manners to capture the image with a camera embedded in the smartphone according to different users. This paper proposes the Check4Urine system, a novel smartphone-based portable urine-analysis tester and provides three techniques to improve such a performance degradation problem robust to various test environments and disturbances, which are the compensation algorithm to correct the varying illumination effect, an urine strip detection algorithm robust to edge loss of the object image, and the color decision algorithm based on the pre-processed reference table. Experimental results show that the proposed Check4Urine system increases the accuracy of urine-analysis by 20-50% at various test conditions, compared with the existing image-processing based mobile urine tester.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2003년도 Proceedings of ACRS 2003 ISRS
• /
• pp.144-146
• /
• 2003

Techniques for rock slope management and assessment must be developed for the prevention and mitigation of rock fall hazards. To enable this, the rock discontinuity such as fault and joint data must be surveyed, analysed and managed. For this, the discontinuities were detected by automatic and semi-automatic method using DEM and ortho-rectified image of rock slope and the rock slope analysis and management system was developed using GIS. Using the system, slope locations and discontinuities data were constructed to spatial database. The system is consist of ‘Data Management’, ‘Rock Slope DB’, ‘Basic Information’, ‘Image Processing’, ‘Image Analys ing’, ‘Edit’, ‘View’, ‘Theme’, ‘Graphic’, ‘Window’ and ‘Help’. The system was developed using avenue of ArcView 3.2.

• Korean Journal of Geomatics
• /
• 제1권1호
• /
• pp.85-93
• /
• 2001

Geospatial database, which is the basis for Geo-Spatial Information Systems, is produced by conventional mapping methods. Recently, with increased demand for digital forms of the geospatial database, studies are carried out to automate its production. The automated mapping system is composed of the image acquisition subsystem, positioning subsystem, point referencing subsystem and the visualization subsystem. The image acquisition subsystem is the most important part of the overall production line because it is the starting point and will affect all subsequent processes. This paper presents a software analysis and design of the image acquisition subsystem. The design was carried out using the Unified Modeling Language which is a modeling method used extensively in the software engineering field.

• 한국광학회지
• /
• 제15권3호
• /
• pp.234-240
• /
• 2004

집적결상법으로 3차원 영상을 재생하는 경우 재생영상의 일부분이 중복 또는 누락되는 왜곡현상이 나타난다. 본 논문에서는 이러한 왜곡현상의 원인을 분석하고 정량적 해석을 시도하였다. 집적결상에서 사용 가능한 몇 가지 기본적인 변수들을 도입하여 왜곡을 정량화하고 최소화될 조건을 조사하였으며, 그 특성을 실험을 동해 확인하였다.

• 한국의류학회지
• /
• 제30권8호
• /
• pp.1275-1287
• /
• 2006

This was to confirm the Web-SFAS and street Fashion trend 2005. Korea women's street fashion in 2005 S/S was analyzed by using the Web-SFAS(Street Fashion Analysis System) that was developed and completed in the earlier research. Total 270 women who were sensitive to the fashion were surveyed around in 5 nationwide cities and 9 commercial zones. The analysis results of the questionnaire and image survey in 2005 S/S for each commercial zone is shown as follows: 1. The sequence of care with most interest in producing fashionable shape is in the order of clothes>accessory>hair style>make up>shoes, and it was found that the proper reason for visiting the commercial zone is to meet with friends than shopping. 2. Most preferred item for tops were shirts and T-shirts, while for bottoms were blue jeans of denim material in various colors and designs. The color had the clear winner in white color with the pink or yellow line of colors preferred in all commercial zones. The blue jeans and skirt items were the most dominant with the preference on the blue color. The color and image were diversely favored with light, natural, modern, elegance and warm images. 3. The semi-formal image was preferred in women's street fashion, 2005 S/S in Apgujeong-dong, Hapseong-dong, Myeong-dong, Hongik University area, Dongseong-ro, and the active comfort casual image was preferred by womens in Seo-myeon, Nampo-dong and Sangnam-dong. However the retro image were represented in Daehak-ro specially. The tendency of street fashion for women had the difference of their own trends for each commercial zone, rather then the regional differences for each city with the conspicuous preference of color for the items.

• 한국정밀공학회지
• /
• 제20권11호
• /
• pp.32-40
• /
• 2003

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2005년도 Proceedings of ISRS 2005
• /
• pp.763-765
• /
• 2005

In the area of remote sensing, an immense number of images are continuously generated by various remote sensing systems. These images must then be managed by a database system efficient storage and retrieval. There are many types of image database systems, among which the content-based image retrieval (CBIR) system is the most advanced. CBIR utilizes the metadata of images including the feature data for indexing and searching images. Therefore, the performance of image retrieval is significantly affected by the storage method of the image metadata. There are many features of images such as color, texture, and shape. We mainly consider the shape feature because shape can be identified in any remote sensing while color does not always necessarily appear in some remote sensing. In this paper, we propose a metadata representation and storage method for image search based on shape features. First, we extend MPEG-7 to describe the shape features which are not defined in the MPEG-7 standard. Second, we design a storage schema for storing images and their metadata in a relational database system. Then, we propose an efficient storage method for managing the shape feature data using a Wavelet technique. Finally, we provide the performance results of our proposed storage method.

• 한국데이터정보과학회:학술대회논문집
• /
• 한국데이터정보과학회 2005년도 추계학술대회
• /
• pp.35-45
• /
• 2005

Data fusion is method to combination data. The purpose of this study is to design and implementation for street fashion information analysis system using data fusion. It can offer variety and actually information because it can fuse image data and survey data for street fashion. Data fusion method exists exact matching method, judgemental matching method, probability matching method, statistical matching method, data linking method, etc. In this study, we use exact matching method. Our system can be visual information analysis of customer's viewpoint because it can analyze both each data and fused data for image data and survey data.

• Journal of Advanced Marine Engineering and Technology
• /
• 제23권1호
• /
• pp.79-87
• /
• 1999

The most part of the faculty drop a trouble and damage of machine equipment even if whatever cause they break out take place at local and trifling place and the factor dominating their trouble is due to wear debris occurred in the lubricated machine moving surface. This study has been car-ried out to identify morphology of wear debris on the lubricated machine moving system by means of computer image analysis. Namely the wear debris contained in lubricating oil extracted from movable machine equipment will be filtered through membrane filter(void diameter 0.45${\mu}m$) and will be analyzed with its data information such as 50% volume diameter aspect roundness and reflectivity. Morphological characteristic of wear debris is easily distinguished by four shape parameters it is necessary to divide small class of every 100 wear debris in total wear particles in order to distinguish morphological characteristic of wear debris more easily by computer image analysis. We are sure that operation condition diagnosis of the lubricated machine moving surfaces is possible by computer image analysis.