• 전자공학회논문지
• /
• 제52권8호
• /
• pp.89-96
• /
• 2015

인간의 시각은 색순응을 통해서 사물의 색을 광원의 색에 영향 없이 인지 할 수 있다. 반면에, 카메라는 입력 값을 그대로 기록하기 때문에, 광원에 따라 물체의 색이 다르게 나타난다. 최근에 희박성 제약조건의 비음수 행렬 분해(nonnegative matrix factorization with sparseness constraint; NMFsc)를 이용한 광원추정 방법이 제안되었다. 이 방법은 낮은 희박성 제약조건을 사용해서 광원을 추정하고, 높은 희박성 제약조건을 사용해서 반사율을 추정한다. 하지만, 희박성 제약조건의 비음수 행렬분해를 이용한 광원 추정 방법은, 영상의 전역적인 정보를 사용하므로, 영상에서 동일한 색이 넓은 영역에 존재하는 경우, 추정된 광원이 큰 오차를 가진다. 이러한 단점을 보완하기 위해, 영상에서 주색도 분석과 희박성 제약조건의 비음수 행렬 분해를 이용한 광원 추정 방법을 제안하였다. 먼저 주색도를 분석하기 위해 영상을 색도 좌표계로 옮기고 색도 히스토그램을 이용하여 유사한 색도를 가지는 영역들로 영상을 분할한다. 다음으로 영상의 주색도는 분할된 영상들 중 색도의 표준편차가 가장 적은 영상의 색도로 선택한다. 마지막으로 주색도 분석 결과와 희박성 제약조건의 비음수 행렬 분해를 이용해 입력 영상에서 주색도 성분을 제거하고 최종적인 광원을 추정한다. 실제 촬영 영상에 대한 평균 각오차를 사용하여 기존의 방법과의 성능을 비교하였고, 그 결과 제안하는 방법의 평균 각 오차는 5.5를 나타내어 영상의 주 색도를 포함하여 광원을 추정한 기존 방법의 평균 각 오차 5.7 보다 우수한 성능을 나타내었다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 추계종합학술대회 논문집(4)
• /
• pp.9-12
• /
• 2000

This paper proposes an improved image illumination estimation method based on the conventional color constancy algorithm. The most important process of color constancy algorithm is the estimation of the spectral distributions of illuminant of an input image. To estimate of the spectral distributions of illuminant of an input image, we use the brightest pixel values and the values of surface reflectance of an input image using a principal component analysis of the given munsell chips. We estimate a CIE tristimulus values of an input image using the estimated .spectral distribution of illuminant and recover an image by scaling it regularity. From the experimental results, the proposed method was effective in estimating the image illumination

• 한국멀티미디어학회논문지
• /
• 제25권8호
• /
• pp.991-998
• /
• 2022

Recently, object recognition using image/video signals is rapidly spreading on autonomous driving and mobile phones. However, the actual input image/video signals are easily exposed to a poor illuminance environment. A recent researches for improving illumination enable to estimate and compensate the illumination parameters. In this study, we propose VE-DCE (video enhancement zero-reference deep curve estimation) to improve the illumination of low-light images. The proposed VE-DCE uses unsupervised learning-based zero-reference deep curve, which is one of the latest among learning based estimation techniques. Experimental results show that the proposed method can achieve the quality of low-light video as well as images compared to the previous method. In addition, it can reduce the computational complexity with respect to the existing method.

• Journal of Electrical Engineering and Technology
• /
• 제10권4호
• /
• pp.1822-1829
• /
• 2015

In this study, artificial neural networks were used to determine the intensity of brightness in interior spaces. The illumination elements to illuminate indoor spaces were considered, not individually, but as a system. So, during the planned maintenance periods of an illumination system, after its design and installation, simple brightness level measurements must be taken. For a three-dimensional evaluation of the brightness level in indoor spaces in a speedy and accurate manner, the obtained brightness level measurement results and artificial neural network model were used. Upon estimation of the most suitable brightness level for indoor spaces by using the artificial neutral network model, the energy demands required by the illumination elements decreased. Consequently, in this study, with estimations of brightness levels, the extent to which the artificial neutral networks become successful was observed and more correct results have been obtained in terms of both economy and usage.

• 디지털산업정보학회논문지
• /
• 제13권1호
• /
• pp.125-133
• /
• 2017

Color in an image is determined by illuminant and surface reflectance. So, to recover unique color of object, estimation of exact illuminant is needed. In this study, the illumination models suggested to get the object color constancy with the physical illumination model based on physical phenomena. Their characteristics and application limits are presented and the necessity of an extended illumination model is suggested to get more appropriate object colors recovered. The extended illumination model should contain an additional term for the ambient light in order to account for spatial variance of illumination in object images. Its necessity is verified through an experiment under simple lighting environment in this study. Finally, a reconstruction method for recovering input images under standard white light illumination is experimented and an useful method for computing object color reflectivity is suggested and experimented which can be induced from combination of the existing illumination models.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제33권1호
• /
• pp.84-94
• /
• 2006

영상의 컬러 정보는 조명 환경에 따라 변한다. 이 경우 영상으로부터 조명 컬러를 추정하고 물체 본래의 컬러를 복원하는 것을 컬러 항상성이라 한다. 본 논문에서는 이색성 반사 모델에 기반한 새로운 컬러 항상성 복원 기법을 제안한다. 이색성 반사 모델에 기반하여 조명 컬러를 추정하기 위해서는 정확한 이색성 선을 추출하는 것이 매우 중요하다. 이를 위해 본 논문에서는 이색성 기울기와 이색성 선 공간을 제안하고, 이를 이용하여 이색성 선을 추정하는 기법을 제안한다. 또한 얻어진 이색성 선을 이용하여 단일 및 다중 컬러 영상에 대한 조명 추정 기법도 제안한다.

• Current Optics and Photonics
• /
• 제4권4호
• /
• pp.317-325
• /
• 2020

Accurate estimation of spatial frequencies and phases for illumination patterns are essential to reconstructing super-resolution images in structured illumination microscopy (SIM). In this manuscript, we propose the improved component cross-correlation (ICC) algorithm, which is based on optimization of the cross-correlation values of the overlapping information between various spectral components. Compared to other algorithms for spatial-frequency and phase determination, the results calculated by the ICC algorithm are more accurate when the modulation depths of the illumination patterns are low. Moreover, the ICC algorithm is able to calculate the spatial frequencies and phases simultaneously. Simulation results indicate that even if the modulation depth is lower than 0.1, the ICC algorithm still estimates the parameters precisely; the images reconstructed by the ICC algorithm are much clearer than those reconstructed by other algorithms. In experiments, our home-built SIM system was used to image bovine pulmonary artery endothelial (BPAE) cells. Drawing support from the ICC algorithm, super-resolution images were reconstructed without artifacts.

• 대한전자공학회논문지SP
• /
• 제38권3호
• /
• pp.292-303
• /
• 2001

임의의 물체색은 장면(scene)에 존재하는 조명과 물체 표면의 특성에 의해 결정되므로, 정확한 물체색을 표현하기 위해서는 조명색의 추정이 중요하다. 본 논문은 인지광원(perceived illumination) 현상을 확장한 방법과, 광휘점(highlight) 방법을 각각 제안하고, 두가지 방법을 결합하는 결합적 조명색 추정방법을 제안한다. 인지광원 방법은 개략적인 해의 범위를 결정하는 면에서는 안정성이 보장되나, 정확성의 측면에서는 입력영상의 내용에 의존적인 경향이 있는 단점이 있다. 광휘점 방법은 입력영상의 내용에 의존적이지 않으며, 정확한 해를 제시하는 장점이 있으나, 최종적인 해를 결정하기 위해 폭넓은 범위를 가지는 교차점인 다수의 후보들을 고려해야 하는 단점이 있다. 그러므로 본 논문에서는 두 가지 방법의 상호보완적인 특성을 이용하여, 인지광원 방법의 추정결과를 가능한 해의 개략적인 범위로 설정하고, 광휘점 방법으로부터 추출된 후보점 및 분포 클러스터(cluster)들의 특성을 고려하여 최종적인 해를 결정하는 알고리즘을 제안한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권11호
• /
• pp.4170-4188
• /
• 2014

In this study, we present a denoising algorithm for high-frame-rate videos in an ultra-low illumination environment on the basis of Kalman filtering model and a new motion segmentation scheme. The Kalman filter removes temporal noise from signals by propagating error covariance statistics. Regarded as the process noise for imaging, motion is important in Kalman filtering. We propose a new motion estimation scheme that is suitable for serious noise. This scheme employs the small motion vector characteristic of high-frame-rate videos. Small changing patches are intentionally neglected because distinguishing details from large-scale noise is difficult and unimportant. Finally, a spatial bilateral filter is used to improve denoising capability in the motion area. Experiments are performed on videos with both synthetic and real noises. Results show that the proposed algorithm outperforms other state-of-the-art methods in both peak signal-to-noise ratio objective evaluation and visual quality.

• 제어로봇시스템학회논문지
• /
• 제22권9호
• /
• pp.738-744
• /
• 2016

In this paper, a robust visual odometry system has been proposed and implemented in an environment with dynamic illumination. Visual odometry is based on stereo images to estimate the distance to an object. It is very difficult to realize a highly accurate and stable estimation because image quality is highly dependent on the illumination, which is a major disadvantage of visual odometry. Therefore, in order to solve the problem of low performance during the feature detection phase that is caused by illumination variations, it is suggested to determine an optimal threshold value in the image binarization and to use an adaptive threshold value for feature detection. A feature point direction and a magnitude of the motion vector that is not uniform are utilized as the features. The performance of feature detection has been improved by the RANSAC algorithm. As a result, the position of a mobile robot has been estimated using the feature points. The experimental results demonstrated that the proposed approach has superior performance against illumination variations.