• Annual Conference of KIPS
• /
• 2022.11a
• /
• pp.643-645
• /
• 2022

많은 온라인 쇼핑몰에서 색상 기반 필터링 서비스나 추천 시스템을 제공하지만, 수동 분류는 많은 시간이 들고 오류 위험이 있다. 본 연구의 실험에서는 먼저 분석할 의류 이미지를 실루엣 분석으로 수행한 경우와 수행하지 않는 경우의 k-평균 군집화 알고리즘으로 가장 우세한 색상 군집의 중심값을 도출하는데, 만약 군집 개수가 2개 이상이면 보다 큰 군집의 중심값만을 고려한다. 이 중심값을 이용해 사전 학습한 k-최근접 이웃 알고리즘으로 색상 클래스를 분류한다. 실험 결과 실루엣 분석을 수행하지 않은 k-평균 군집화 알고리즘을 사용한 분류 방식이 정확도와 수행 시간 모두 매우 준수하였으나, 배경색이 존재하여 의류 색 분석에 영향을 줄 수 있는 경우 잘못 분류한다는 문제도 있다.

• Annual Conference of KIPS
• /
• 2024.05a
• /
• pp.440-443
• /
• 2024

한의학의 증상을 파악하는 한의 변증 진단에서 혀의 형태와 색상을 파악하는 것에서 설질과 설태의 색상을 파악하는 것이 핵심이다. 혀 촬영 장치로 촬영하더라도 조명의 빛 종류와 빛 강도, 위치에 따라 컬러가 변환됨에 따라 혀 취득장치의 컬러교정이 필요하다. 컬러교정을 하지만, 오버피팅, 측정 오류와 노이즈 또는 색상 공간의 불균형으로 컬러의 왜곡이 발생할 수 있다. 이 연구에서는 24개의 patch 중에서 오류를 최소화할 수 있는 patch를 선정하는 최적화의 방법을 제시하려 하였다. 24개 patch 중 몇 개의 선택한 patch의 평균값을 구하여, 기준 값으로 변환시키는 변환 행렬을 구하고, 교정 값을 구하고, 교정된 24 patch 값의 평균과 기준 값과의 오차를 구하여 최소가 되는 변환 행렬을 구하였다. 이 방법은 컬러 챠트의 개수를 줄여 장치의 부피를 줄이는데 활용될 수 있다. 또한 이 방법은 외부의 빛이 차단된 촬영장치에서 활용되었지만, 조명조건이 안정된 개방된 공간에서도 스마트폰을 이용하여 촬영하는데 활용 가능할 것으로 예상된다.

• Journal of the HCI Society of Korea
• /
• v.2 no.1
• /
• pp.49-56
• /
• 2007

A nonintrusive projection-based AR approach using complementary pattern has been recently proposed and applied to virtual studio. However, the approach faces the tradeoff between the pattern imperceptibility and compensation accuracy. To alleviate the tradeoff, we propose a content adaptive pattern concealment approach. The projector input images (AR images) are divided into rectangular regions and spatial variation and color distribution are computed in the regions. Based on the spatial variation and color distribution, we embed locally different strength of pattern images into different color channels. It is demonstrated that the proposed approach has two opposite advantages by comparing it with the previous (non-adaptive) approach through a variety of experiments and subjective evaluation. Our content adaptive approach can obtain the same performance using weaker pattern than the previous approach and thus significantly improve the imperceptibility of the pattern. On the contrary, our content adaptive approach can make strong pattern less perceptible and thus produce better compensation results.

• Progress in Medical Physics
• /
• v.23 no.3
• /
• pp.199-208
• /
• 2012

The computational human phantom including major radiation sensitive organs at risk (OARs) can be used in the field of radiotherapy, such as the variation of secondary cancer risks caused by the radiation therapy and the effective dose evaluation in diagnostic radiology. The present study developed a Korean adult female voxel phantom, VKH-Woman, based on serially sectioned color slice images of Korean female cadaver. The height and weight of the developed female voxel phantom are 160 cm and 52.72 kg, respectively that are virtually close to those of reference Korean female (161 cm and 54 kg). The female phantom consists of a total of 39 organs, including 27 organs recommended in the ICRP 103 publication for the effective dose calculations. The female phantom composes of $261{\times}109{\times}825$ voxels (=23,470,425 voxels) and the voxel resolution is $1.976{\times}1.976{\times}2.0619mm^3$ in the x, y, and z directions. The VHK-Woman is provided as both ASCII and Binary data formats to be conveniently implemented in Monte Carlo codes.

• Korean Journal of Remote Sensing
• /
• v.32 no.6
• /
• pp.639-645
• /
• 2016

Yeongjong-do is seventh largest island in the west coast of Korea which is 4.8 km away in the direction of south-west from Incheon. The mudflat area around the Yeongjong-do has variable dimension according to tide level. In addition, Yeongjong-do is important area with high environmental value as wintering sites for migratory birds. The mudflat of Yeongjong-do is also meaningful region because it is used as place of education and tourist attraction. But, there are increasing concerns about preservation of mudflat area caused by artificial development such as land reclamation project and Incheon airport construction. In this paper, mudflat area was detected using Landsat 7 ETM+ images that United States Geological Survey (USGS) is providing the data in 16 days period. The false color composite was made from band 7, 5, and 3 that could dividing boundary between water and land for the purpose of appearance of boundary line in mudflat region. This area was calculated using results of classification based on false color composite images and was digitized through repetitive algorithm during research of period. Therefore, the change of northeastern area in Yeongjong-do was detected according to tide level during 16 years from 2000 to 2015 on the basis of providing period at tide station. This paper will expect as indicator for range of area in same tide level prior to the start of the research for preservation of mudflat. It will be also scientific grounds about change of mudflat area caused by artificial development.

• Archives of design research
• /
• v.17 no.4
• /
• pp.233-240
• /
• 2004

Living in the modern age with well-developed scientific technologies, all of us are enjoying convenient lives because of 'energy'. Korea, poor in resources, is importing a large portion of its energy sources from abroad but energy consumption shows an upward tendency due to the continuing economic growth and the improvement of living conditions. The atomic energy is considered a self-reliant, alternative energy source like our country. However, it is necessary to educate the people on and publicize atomic power generation in the face of the widespread negative recognition that the atomic power plant is a hazardous facility. The study approaches to these matters with a human-friendly and environment-friendly coloring plan in the perspective of environment coloring plan. The study aims to minimize negative images of the atomic power, while highlighting its friendly and positive images so as to enhance the confidence of the people on the atomic power and to create a clean image for the atomic power. For this goal, the study examined and analyzed cases of Japanese nuclear power plants and domestic nuclear power plants, and also carried out an on-site survey in the sites in which nuclear power plants would be constructed to extract concrete colors through the analyses of their natural environment and actual conditions. The study also carried out a survey of residents in the regions to induce their participation, and reflected the survey results to the coloring plan. The study is expected to create a stable and friendly image of the nuclear power plant through materializing its environment-friendly image and remove negative recognition that the people have on the nuclear power plant. It also attempted an external environment-coloring plan a s a strategic means for positive publicity and through this, is expected to ultimately contribute to the creation of the new images of nuclear ower plants.

• Journal of Broadcast Engineering
• /
• v.26 no.3
• /
• pp.283-295
• /
• 2021

This paper proposes a method for generating dense depth map using information of color images and depth map generated based on lidar based on self-organizing map. The proposed depth map upsampling method consists of an initial depth prediction step for an area that has not been acquired from LiDAR and an initial depth filtering step. In the initial depth prediction step, stereo matching is performed on two color images to predict an initial depth value. In the depth map filtering step, in order to reduce the error of the predicted initial depth value, a self-organizing map technique is performed on the predicted depth pixel by using the measured depth pixel around the predicted depth pixel. In the process of self-organization map, a weight is determined according to a difference between a distance between a predicted depth pixel and an measured depth pixel and a color value corresponding to each pixel. In this paper, we compared the proposed method with the bilateral filter and k-nearest neighbor widely used as a depth map upsampling method for performance comparison. Compared to the bilateral filter and the k-nearest neighbor, the proposed method reduced by about 6.4% and 8.6% in terms of MAE, and about 10.8% and 14.3% in terms of RMSE.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.7 no.3
• /
• pp.383-391
• /
• 2017

In image processing, image segmentation has been studied by various methods in a long time. Image segmentation is the process of partitioning a digital image into multiple objects and face detection is a typical image segmentation field being used in a variety of applications that identifies human faces in digital images. In this paper, we propose a method for extracting the contours of faces included in images. Using the Viola-Jones algorithm, to do this, we detect the approximate locations of faces from images. But, the Viola-Jones algorithm could detected the approximate location of face not the correct position. In order to extract a more accurate face region from image, we use skin color in this paper. In details, face region would be extracted using the analysis of horizontal and vertical histograms on the skin area. Finally, the face contour is extracted using snake algorithm for the extracted face area. In this paperr, a modified snake energy function is proposed for face contour extraction based snake algorithm proposed by Williams et al.[7]

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2007.04a
• /
• pp.141-144
• /
• 2007

Satellite image for national land monitoring is required high resolution and natural color with multi spectral band. the image is expensive as higher resolution. We need cheap image relatively in economic viewpoint but the image serves sufficient resolution to monitor national land. We merged two images to one image and evaluated the result. the two images which are used at the merge test are high resolution(2.5m per pixel) panchromatic and low resolution(10m per pixel) multi spectral image of SPOT-5 satellite. The result of this study. We made the merge image to have sufficient resolution for national monitoring.

• Smart Media Journal
• /
• v.2 no.2
• /
• pp.14-19
• /
• 2013

Recently, due to the increase of the use of scanned images, the text segmentation techniques, which play critical role to optimize the quality of the scanned images, are required to be updated and advanced. In this study, an algorithm has been developed based on the modification of Gaussian mixture model (GMM) by integrating the calculation of Gaussian detection gradient and the estimation of the number clusters. The experimental results show an efficient method for text segmentation in natural scenes such as storefronts, street signs, scanned journals and newspapers at different size, shape or color of texts in condition of lighting changes and complex background. These indicate that our model algorithm and research approach can address various issues, which are still limitations of other senior algorithms and methods.