• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.04a
• /
• pp.938-940
• /
• 2017

본 논문에서는 화재의 조기 감지를 위하여 카메라 입력 영상으로부터 화염을 검출하는 알고리즘을 제안한다. 화염은 특정 RGB 좌표계를 가지며 지속해서 형태가 변화하며 움직인다. 제안하는 화염 검출 알고리즘은 먼저 야외 환경에서 조도의 변화에 관계없이 화염 검출 알고리즘을 적용하기 위해 Color Constancy 알고리즘을 적용한다. 그 후 화염의 RGB 좌표계와 움직임의 변화를 측정하여 후보영역을 설정하고 Opponent SURF와 SVM을 통해 최종 화염을 검출한다. 컴퓨터 시뮬레이션을 통하여 제안하는 알고리즘으로 화염을 검출할 수 있음을 확인하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2013.06a
• /
• pp.258-259
• /
• 2013

광원 및 조명이 미약한 환경에서 획득된 저조도 영상은 인지적 및 색 왜곡적 측면에서 취약점을 가진다. 영상의 색 복원을 위한 연구인 색 항등성 기법은 저조도 환경에 적합하지 않기 때문에 저조도 영상을 대상으로 적용할 경우에는 좋은 성능을 내지 못한다. 이러한 문제를 해결하기 위하여 본 논문에서는 저조도 영상의 색 복원을 위한 톤 매핑 및 색 항등성 기법에 대해 분석한다. 톤 매핑 기법은 저조도 영상의 밝기를 개선해 색 항등성 기법의 적용을 가능하도록 하기 위해 사용되며, 이후 다양한 색 항등성 기법을 밝기 조절된 저조도 영상에 적용해 색 복원에 적합 여부를 판단한다.

• Proceedings of the KAIS Fall Conference
• /
• 2011.12b
• /
• pp.683-685
• /
• 2011

저조도 환경에서 획득된 영상은 대부분의 픽셀이 낮은 RGB 값을 가지기 때문에 물체가 가지는 색의 식별 및 물체 간의 구별이 어렵다는 문제점을 갖는다. 이러한 문제는 이론적으로 영상 내 존재하는 광원의 영향을 제거하는 것을 목적으로 하는 색 항등성 기법을 적용하여 해결이 가능하다. 저조도 영상에 적합한 색 항등성 기법을 찾기 위하여 본 논문에서는 Barnard 데이터 셋을 바탕으로 하는 저조도 합성 영상을 생성하고 이를 기반으로 다양한 색 항등성 기법을 평가한다. 저조도 합성 영상은 원하는 장면을 가지는 영상과 GTD를 생성할 수 있는 장점이 있기 때문에 실험 영상으로 사용된다. 성능 평가는 색 항등성 기법을 적용한 결과 영상과 GTD 영상을 비교하여 수행된다.

• Proceedings of the KAIS Fall Conference
• /
• 2008.11a
• /
• pp.320-322
• /
• 2008

본 논문에서는 조명의 변화에 강건한 얼굴 검출 알고리즘을 제안한다. 영상내의 조명 성분을 줄이기 위하여 컬러 일관성(color constancy) 알고리즘 중 Simplified Horn 기법을 적용한 후 색 정보를 이용하여 얼굴 후보영역을 결정한다. 이렇게 결정된 얼굴 후보영역 중 얼굴영역과 헤어영역의 여러 기하학적인 정보를 이용하여 실제 얼굴 영역을 판단한다. 제안한 알고리즘은 다양한 조명 성분을 갖는 여러 영상에서 테스트 되었으며 높은 검출률을 보였다.

• Journal of Korea Multimedia Society
• /
• v.4 no.2
• /
• pp.97-107
• /
• 2001

In this paper, an illuminant estimation algorithm of a chromatic digital images proposed. The proposed illumination estimation method has two phases. First, the surface spectral reflectances are recovered. In this case, the surface spectral reflectances recovered are limited to the maximum highlight region (MHR) which is the most achromatic and highly bright region of an image after applying intermediate color constancy process using a modified gray world algorithm. Next, the surface reflectances of the maximum highlight region are estimated using the principal component analysis method along with a set of given Munsell samples. Second, the spectral distribution of reflected lights of MHR is selected from the spectral database. That is a color difference is compared between the reflected lights of the MHR and the spectral database that is the set of reflected lights built by the given Munsell samples and a set of illuminants. Then the closest colors from the spectral database are selected. Finally, the illuminant of an image can be calculated dividing the average spectral distributions of reflected lights of MHR by the average surface reflectances of the MHR. In order to evaluate the proposed algorithm, experiments with artificial and real captured color-biased scenes were performed and numerical comparison examined. The proposed method was effective in estimating the spectral of the given illuminant sunder various illuminants.

• Korean Journal of Cognitive Science
• /
• v.25 no.1
• /
• pp.51-71
• /
• 2014

This paper presents an image enhancement method in real road traffic scenes. The images captured by the camera on the car cannot keep the color constancy as illumination or weather changes. In the real environment, these problems are more worse at back light conditions and at night that make more difficult to the applications of the vision-based intelligent vehicles. Using the existing image enhancement methods without considering the position and intensity of the light source and their geometric relations the image quality can even be deteriorated. Thus, this paper presents a fast and effective method for image enhancement resembling human cognitive system which consists of 1) image preprocessing, 2) color-contrast evaluation, 3) alpha blending of over/under estimated image and preprocessed image. An input image is first preprocessed by gamma correction, and then enhanced by an Automatic Color Enhancement(ACE) method. Finally, the preprocessed image and the ACE image are blended to improve image visibility. The proposed method shows drastically enhanced results visually, and improves the performance in traffic sign detection of the vision based intelligent vehicle applications.

• Journal of Broadcast Engineering
• /
• v.18 no.1
• /
• pp.21-30
• /
• 2013

It is a challenging problem to most of the image processing when the light source is unknown. The color of the light source must be estimated in order to compensate color changes. To estimate the color of the light source, additional assumption is need, so that we assumed color distribution according to the light source. If the pixels, which do not satisfy the assumption, are used, the estimation fails to provide an accurate result. The most popular color distribution assumption is Grey-World Assumption (GWA); it is the assumption that the color in each scene, the surface reflectance averages to gray or achromatic color over the entire images. In this paper, we analyze the characteristics of the camera response function, and the effect of the Grey-World Assumption on the pixel value and chromaticity, based on the inherent characteristics of the light source. Besides, we propose a novel method that detects important pixels for the color estimation of the light source. In our method, we firstly proposed a method that gives weights to pixels satisfying the assumption. Then, we proposed a pixel detection method, which we modified max-RGB method, to apply on the weighted pixels. Maximum weighted pixels in the column direction and row direction in one channel are detected. The performance of our method is verified through demonstrations in several real scenes. Proposed method better accurately estimate the color of the light than previous methods.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.11
• /
• pp.457-464
• /
• 2021

The advent of the big data era has enabled the rapid development of deep learning that learns rules by itself from data. In particular, the performance of CNN algorithms has reached the level of self-adjusting the source data itself. However, the existing image processing method only deals with the image data itself, and does not sufficiently consider the heterogeneous environment in which the image is generated. Images generated in a heterogeneous environment may have the same information, but their features may be expressed differently depending on the photographing environment. This means that not only the different environmental information of each image but also the same information are represented by different features, which may degrade the performance of the image analysis model. Therefore, in this paper, we propose a method to improve the performance of the image color constancy model based on Adversarial Learning that uses image data generated in a heterogeneous environment simultaneously. Specifically, the proposed methodology operates with the interaction of the 'Domain Discriminator' that predicts the environment in which the image was taken and the 'Illumination Estimator' that predicts the lighting value. As a result of conducting an experiment on 7,022 images taken in heterogeneous environments to evaluate the performance of the proposed methodology, the proposed methodology showed superior performance in terms of Angular Error compared to the existing methods.

• Journal of Korean Society for Quality Management
• /
• v.44 no.3
• /
• pp.527-540
• /
• 2016

Purpose: The purpose of this study was to propose contrast metric which is based on the human visual perception and thus it can be used to improve the quality of digital images in many applications. Methods: Previous literatures are surveyed, and then the proposed method is modeled based on Human Visual System(HVS) such as multiscale property of the contrast sensitivity function (CSF), contrast constancy property (suprathreshold), color channel property. Furthermore, experiments using digital images are shown to prove the effectiveness of the method. Results: The results of this study are as follows; regarding the proposed contrast measure of complex images, it was found by experiments that HVS follows relatively well compared to the previous contrast measurement. Conclusion: This study shows the effectiveness on how to measure the contrast of complex images which follows human perception better than other methods.

• Journal of Broadcast Engineering
• /
• v.19 no.1
• /
• pp.75-82
• /
• 2014

This paper proposes an illuminant compensation method using a camera noise analysis without segmentation in the dichromatic reflectance model. In general, pixels within highlight regions include large amounts of information on the image illuminant. Thus, the analysis of highlight regions provides a relatively easy means of determining the characteristics of an image illuminant. Currently, conventional methods require regional segmentation and the accuracy of this segmentation then affects the illuminant estimation. Therefore, the proposed method estimates the illuminant without segmentation based on a conformity assessment of highlight regions. Furthermore, error factors, such as noise and sensor non-uniformity, can be reduced by the conformity assessment.