Recovery of Surface Reflectances Using Video Camera Image Signals under Unknown Illuminations

Estimation of surface spectral reflectance functions and colorimetric values of object colors was carried out for the image data captured by video camera under unknown illuminations. The estimation method is based on the finite-dimensional linear model of illuminants and surfaces using a single reference reflectance with known surface spectral reflectance in a scene. Experimental results showed that recovered surface spectral reflectance functions are the almost same for different illuminations. This method is useful for the colorimetric calibration of the devices, since any set of orthonormal basis vectors to represent illuminations and any set of eigen vectors derived from principal component analysis on the data set of a color chart to express surfaces of different objects can be applied to the estimation. The recovered surface reflectances were evaluated by color differences in CIE L*a*b* color space.     

Illuminant Invariant Recovery of Surface Reflectances and Estimation of Colorimetric Values from Image Signals

Computer simulation based on the finite-dimensional linear model for computational color constancy model has been carried out to estimate colorimetric values using image signals. We estimated the colorimetric values placing a reference object with known spectral reflectance in a scene under an unknown illuminant. It was found that prior knowledge of illuminants and surfaces of objects in a scene is not required for the estimation: any set of basis vectors to express the illuminants and any set of eigen vectors obtained by principal component analysis on data set of a color chart to represent the surfaces of different objects can be applied to the estimation.     

Total Colorimetric Evaluation of a Set of Color Sensors for a Variety of Illuminants and Objects

Recently, a new colorimetric evaluation model for a set of color sensors was proposed by the author (N. Shimano: J. Inst. Image Electron. Eng. Jpn. 29 (2000) 506 and 517.), and the quality was shown to be defined by the ratio of the statistical mean energy of color stimuli at the retina (SMECS) to the captured energy of the SMECS by the sensors. It was also shown that the quality depends on the illuminants and objects. The experimental results agreed quite well with the proposed model. However, an extension of the model is needed to evaluate the full colorimetric quality of an image acquisition device for a variety of illuminants and objects. In this paper the extended model called the total colorimetric quality model and experimental results for sensors under such a variety are presented.     

Measurement of Color Constancy by Color Memory Matching

Degree of color constancy was measured when color memory was involved in color comparison judgment. We used the Optical Society of America (OSA) Uniform Color Scales as stimulus color samples, and chose 20 color samples as test stimuli. Four illuminants of 1700, 3000, 6500, and 30, 000 K were tested. The observer, completely adapted to a test illuminant, saw a test color sample and stored its color in his memory. After being readapted to the reference white (6500 K), he started selecting a color sample from among the 424 OSA samples which matched the test sample in his memory. We employed a memory matching method called cascade color matching, in which the number of selected color-samples was gradually reduced in four stages. In the final stage, the observer selected a color sample. The results show that, for most test colors, the distributions of selected colors in stages 1 to 4 were similar among all illuminants, and that the u’v’ chromaticity distance between a test color under 6500 K and its matched color was quite short. These indicate that good color constancy was retained in memory color comparison.     

Color Constancy in a Photograph Perceived as a Three-Dimensional Space

It is known that color constancy does not hold in a photograph. This could be because the photograph is recognized as a two-dimensional paper. Based on the concept of the recognized visual space of illumination (RVSI), it is predicted that color constancy holds in the photograph if it is perceived as a 3-D scene. We examined whether the color constancy held under a special viewing condition. A photograph of a room under incandescent illumination was shown under daylight illumination. We tested the neutral color perception of a stimulus on the photograph both with and without a dimension-up viewing box showing the photograph alone monocularly. The results showed good color constancy when a subject observed the photograph with the viewing box. It was also shown that the degree of color constancy decreased for a jumbled photograph without 3-D information. Our results suggest that the recognition of a space and illumination are important in color perception.     

Color Enhancement in Multispectral Image Using the Karhunen-Loeve Transform

We propose a new method for the color enhancement of multispectral image in the visible wavelength region. The purpose of the proposed method is to explore the weak features contained in a specific wavelength by discounting the major color distribution. Such examination will be valuable in visual inspection applications, for example, a medical examination using color image to find a small spectral change of an abnormal part. In this method, Karhunen-Loeve (KL) transform is applied to multispectral data, and specific wavelength components of only high-order KL coefficients are amplified while low-order coefficients are not changed to retain the major color distribution. In the experiment, this method was applied to multispectral images: a printed test image and a human skin image of a bruised arm were captured by a 16-band multispectral camera. The resultant images were compared with the images obtained by saturation enhancement and that obtained by applying the proposed method to the 3-band image. The method successfully visualized the features, which are almost invisible in natural color images, with less change in background color than saturation enhancement.     

Suppression of Noise Effects in Color Correction by Spectral Sensitivities of Image Sensors

The accurate transformation of device responses to tristimulus values is important at an early stage in color management and this transformation is usually called color correction. In this paper it is shown that the influence of noise on color correction can be suppressed if the square of the singular values by the singular value decomposition (SVD) of a sensor matrix represented by surface reflectance, SLVA^1/2, is larger than the noise variance, where S is the M × N matrix of M channel spectral sensitivities, L is the N × N diagonal matrix for an recording illuminant, V is the N × N matrix which is composed of eigenvectors of an autocorrelation matrix for spectral reflectance of samples, A is the N × N diagonal matrix with eigenvalues of the autocorrelation matrix along the diagonal and N is the number of the dimension, respectively. A method to suppress the influence of noise on color correction by spectral sensitivities of sensors is presented for the first time.     

传统阴极射线管色度预测模型的相干误差分析

传统阴极射线管(CRT)色度预测模型大都建立在三刺激值的可加性原理之上。但是在实际应用中由于红绿蓝(RGB)三个通道之间存在相互干扰, 使得红绿蓝的颜色合成结果在一定程度上偏离了叠加原理, 因此传统阴极射线管色度预测模型必然产生相干误差。通过实验表明, 红绿蓝三个通道之间的干扰主要表现为一种相互抑制作用, 因此造成的传统阴极射线管色度预测模型的预测色差随显示亮度的增加而加大     

Color Modification of Pictures Requiring Same Color Impression as Real Scene

Color constancy is an important feature of the human visual system. Our visual system adapts almost instantly to new illumination and we can perceive the color of objects in a new space correctly whatever color the illumination may be. This feature is not solely that of a commercially available camera. The difference between the human visual system and a camera raises a problem in our observation of pictures. We cannot see the exact color that a photographer intended to show us through a picture. Some color modification must be made to the picture to resolve this. In the present paper, the degree of modification was determined for various color combinations of illumination for taking a picture, FL_L and observing the picture, FL_o. An entire experiment was carried out in an actual room based on the belief that color appearance of objects is determined in relation to the recognized visual space of illumination (RVSI). A subject observed the color of a room lit by FL_L and pictures of the room lit by various colors of FL_L were taken. The pictures were displayed in another room lit by FL_o and the subject chose the picture that gave him/her the same color impression as the room of FL_L. In the case of picture slide, the color of the picture must be modified as if it were taken with the illumination inbetween the color of FL_L and FL_o. In the case of picture print, a similar modification was needed for reddish FL_L and white FL_O, but it was not needed for white FL_L and reddish FL_o.     

Color Conversion Method for Multiprimary Display Using Matrix Switching

The range of reproducible color, i.e., color gamut, of conventional three-primary display devices is sometimes insufficient for reproducing the natural color of an object through color imaging systems. A multiprimary display is being developed for the purpose of reproducing an expanded color gamut using more than three primary colors. In this paper, a color conversion method is proposed to reproduce the natural color by additive mixture of multiprimary colors. The multiprimary color signal in this conversion method is calculated from the three-dimensional color coordinates CIE-XYZ, considering the dynamic range of each primary color. Divided into some linear elements from the polyhedral color solid of the multiprimary display, the conversion considering the constraint of the dynamic range can be performed by simple calculation without iterative calculation or a huge three-dimensional look up table. A fast computation method with a two-dimensional look-up-table is also presented. Using the proposed method, the result of the color reproduction is experimentally demonstrated by a six-primary projection display.     

一种多视点视频自动颜色校正系统

针对多视点视频系统中视点间图像颜色不一致的问题,提出了一种多视点视频自动颜色校正系统。通过求取目标图像和源图像间的颜色校正矩阵,判断其是否满足全局校正的要求;对不满足要求的图像,通过图像分割和K-L变换(Karhunen-Loeve transform),建立起目标图像和源图像中各分割区域间的局部映射关系,并通过感兴趣区域匹配,来实现对源图像的校正,最后通过视频跟踪技术实现对视频图像的校正。以标准的多视点测试图像集为例,通过将新方法与直方图匹配、全局一维线性校正算法等进行比较,表明新方法能消除匹配失真的影响,且具有较好的颜色校正效果。研究结果表明该系统可以很好地揭示图像间的颜色变化关系,并且具有很好的内容自适应性,是一种有效的多视点视频图像系统颜色校正方法。     

孟塞尔色序系统与CIE1931标准色度系统转换新算法

孟塞尔色序系统到CIE标准色度系统的自动转换在颜色和视觉领域是一个很重要的课题.提出了一种基于等色调线和等彩度线插值的新算法,实现孟塞尔色序系统与CIE1931标准色度系统之间颜色的相互转换.实验测试表明,由孟塞尔系统到CIE色度系统的正向转换准确度为ΔY=0.004,Δx=0.000 13,Δy=0.000 13,由CIE色度系统到孟塞尔系统的反向转换准确度达到了ΔV ≈0,ΔH=0.017 5,ΔC=0.013.     

一种基于亮度直方图的自动曝光控制方法

自动曝光控制是现代相机和摄像机必备功能之一。提出一种基于亮度直方图的自动曝光控制算法。基于人类视觉系统(HVS)的视觉注意机制,可认为直方图中大而陡峭的峰值区域对应于图像中的不感兴趣区域。算法寻找直方图中最大的两个峰值区域,并依据峰值区域的大小确定它们中像素亮度的加权值,从而计算图像的加权亮度均值。像素亮度的加权值与其所属峰值区域的大小之间的关系由一组二次型曲线描述,二次型曲线的参量是由图像的背景亮度决定,其中背景亮度由自动曝光系统参量计算得到。在实际应用中采用了模糊逻辑确定最大的两个峰值区域的加权值。实验结果表明该算法对各种场景均能进行有效的自动曝光控制。     

Multispectral Image Compression for Improvement of Colorimetric and Spectral Reproducibility by Nonlinear Spectral Transform

The article proposes a multispectral image compression scheme using nonlinear spectral transform for better colorimetric and spectral reproducibility. In the method, we show the reduction of colorimetric error under a defined viewing illuminant and also that spectral accuracy can be improved simultaneously using a nonlinear spectral transform called Labplus, which takes into account the nonlinearity of human color vision. Moreover, we show that the addition of diagonal matrices to Labplus can further preserve the spectral accuracy and has a generalized effect of improving the colorimetric accuracy under other viewing illuminants than the defined one. Finally, we discuss the usage of the first-order Markov model to form the analysis vectors for the higher order channels in Labplus to reduce the computational complexity. We implement a multispectral image compression system that integrates Labplus with JPEG2000 for high colorimetric and spectral reproducibility. Experimental results for a 16-band multispectral image show the effectiveness of the proposed scheme.     

Color Property of the Recognized Visual Space of Illumination Controlled by Interior Color as the Initial Visual Information

It was shown that the color property of the recognized visual space of illumination, RVSI was controlled by changing the initial visual information by arranging objects in the room all shifting toward orange direction. We constructed two miniature rooms, D and I, both illuminated by the same daylight type fluorescent lamps but arranged with furniture of different color, those in room I shifting toward color as if they were illuminated by an incandescent lamp. Subjects felt as if room I were illuminated by an incandescent lamp. A test patch was placed midair in each room and its color was judged. When the test patches were placed in room I their colors were all perceived to be shifted toward greenish blue compared to those of test patches placed in room D, in spite of having the same illumination. The results imply that the apparent color of an object is determined not by its chromaticity, but in relation to the color property of the RVSI of the room where the object is observed.     

Dynamics of Asymmetric Color Matching

It is known that the human color constancy is not complete. We conducted asymmetric color matching experiments to clarify a simple question: what was it that the observer matched. Observers made apparent-color matches between Munsell color chips under D₆₅ illumination and a color chip presented on a CRT. The observers’ matches showed incomplete color constancy. We applied relative cone-weight transformation to each observer’s results under control condition, so as to equate the unique-white point for each illuminant condition. The result of this simple transformation showed good match to the actual data. Our results suggest what the observer was doing during the asymmetric color matches was picking up apparent-color signal, achieved from the test color chip under D₆₅, and reconstructing it with respect to the unique-white by a simple cone response scaling.     

Color Appearance Explained,Predicted and Confirmed by the Concept of Recognized Visual Space of Illumination

Ten years have passed since we proposed a new concept called recognized visual space of illumination (RVSI). The central idea of the concept assumes that our brain first recognizes how a space is illuminated and then judges colors of anything seen in the space in relation to the RVSI constructed for the space. In another expression we say that the space is recognized first and the color perception next. In this paper some of our experiments that proved the appropriateness of the concept will be introduced. When a white paper was seen through a colored filter we could perceive the paper as white at the same instant we recognized a space through and beyond the filter. When an achromatic patch independent from the room illumination was observed under colored illumination its appearance immediately changed to that roughly opponent to the illumination color. When two gray patches of the same lightness were drawn in a picture of a white grating on a black background on a way that one appeared to locate on this side of the grating and the other in the other side, the former appeared brighter. These all confirmed the predictions based on the RVSI concept.     

LCD液晶显示器的颜色特性研究

定量描述了液晶显示器的颜色特性以及对显示彩色图像的影响。在实验测量的基础上,得到了LCD液晶显示器的颜色特性数据。以3D LUT方法建立了设备颜色空间RGB与标准颜色空间L a b 之间的近似转换模型。通过分析比较它们的色域边界特性,了解了与CRT显示器颜色色域的差异以及影响显示彩色图像时颜色失真的特点。     

近景测量系统全景彩色图像光谱法色度校正

在三维多模式扫描近景测量系统中,使用彩色线阵CCD(电荷藕合器件)可得到景物的一维线状彩色信息,将连续获得的不同角度上的一维彩色图像拼接起来即可形成全景彩色图像。直接在电脑彩色CRT(cathode ray tube, 阴极射线管)显示器上输出该图像,将发生色度失真,需要进行色度校正。通过光谱法色度校正,使得在电脑彩色CRT显示器上输出的图像和原始景物尽可能地保持色度不失真。     

Color appearance seen through a colored filter of various sizes

It will provide us an effective method to study the color perception of the elderly if we can employ for young subjects a pair of glasses with color property chosen to simulate that of the elderly. One problem has to be solved before using such glasses based on the concept of the recognized visual space of illumination, that is, the size of the glasses. When a piece of a white paper is seen through a yellow filter in a room, it naturally appears yellowish. But if the observer uses the filter to cover his/her eyes entirely the white paper returns to its original white; this is known as color constancy. Between these two extreme cases, the color is expected to change from yellow to white. Chromaticness and hue of an achromatic test chart of a small or large size were determined by the elementary color naming method when the chart was seen monocularly through a red, yellow, green or blue filter, respectively, placed at various distances from the eyes. Chromaticness was large at around 40% when the visual field through the filter was equal to or smaller than the test chart, but it rapidly decreased when the visual field extended over the test chart so that objects other than the test chart were included in the visual field. It decreased to almost zero when the filter was about 5 cm or less away from the eye. It was concluded that we do not necessarily have to use goggles to cover the entire visual field but can use a normal pair of glasses made of colored filters for young subjects to study the color perception of the elderly with understanding that the study is only for the color perception experienced while wearing the glasses.