An adaptive color similarity function suitable for image segmentation and its numerical evaluation

In this article, we present an adaptive color similarity function defined in a modified hue‐saturation‐intensity color space, which can be used directly as a metric to obtain pixel‐wise segmentation of color images among other applications. The color information of every pixel is integrated as a unit by an adaptive similarity function thus avoiding color information scattering. As a direct application we present an efficient interactive, supervised color segmentation method with linear complexity respect to the number of pixels of the input image. The process has three steps: (1) Manual selection of few pixels in a sample of the color to be segmented. (2) Automatic generation of the so called color similarity image (CSI), which is a gray level image with all the gray level tonalities associated with the selected color. (3) Automatic threshold of the CSI to obtain the final segmentation. The proposed technique is direct, simple and computationally inexpensive. The evaluation of the efficiency of the color segmentation method is presented showing good performance in all cases of study. A comparative study is made between the behavior of the proposed method and two comparable segmentation techniques in color images using (1) the Euclidean metric of the a* and b* color channels rejecting L* and (2) a probabilistic approach on a* and b* in the CIE L*a*b* color space. Our testing system can be used either to explore the behavior of a similarity function (or metric) in different color spaces or to explore different metrics (or similarity functions) in the same color space. It was obtained from the results that the color parameters a* and b* are not independent of the luminance parameter L* as one might initially assume in the CIE L*a*b* color space. We show that our solution improves the quality of the proposed color segmentation technique and its quick result is significant with respect to other solutions found in the literature. The method also gives a good performance in low chromaticity, gray level and low contrast images. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 156–172, 2017     

Extending the color gamut and dynamic range of an sRGB image using a residual image

Image sources, such as digital camera captures and photographic negatives, typically have more information than can be reproduced on a photographic print or a video display. The information that is lost during the tone/color rendering process relates to both the extended dynamic range and color gamut of the original scene. In conventional photographic systems, most of this additional information is archived on the photographic negative and can be accessed by adjusting the way the negative is printed. However, most digital imaging systems have traditionally archived only a rendered video RGB image. As a result, it is not possible to make the same sorts of image manipulations that historically have been possible with conventional photographic systems. This suggests that there would be an advantage to storing images using an extended dynamic range/color gamut color encoding. However, because of file compatibility issues, digital imaging systems that store images using color encoding other than a standard video RGB representation (e.g., sRGB) would be significantly disadvantaged in the marketplace. In this article, we describe a solution that has been developed to maintain compatibility with existing file formats and software applications, while simultaneously retaining the extended dynamic range and color gamut information associated with the original scenes. With this approach, the input raw digital camera image or film scan is first transformed to the scene‐referred ERIMM RGB color encoding. Next, a rendered sRGB image is formed in the usual way and stored in a conventional image file (e.g., a standard JPEG file). A residual image representing the difference between the original extended dynamic range image and the final rendered image is formed and stored in the image file using proprietary metadata tags. This provides a mechanism for archiving the extended dynamic range/color gamut information, which is normally discarded during the rendering process, without sacrificing interoperability. Appropriately enabled applications can decode the residual image metadata and use it to reconstruct the ERIMM RGB image, whereas applications that are not aware of the metadata will ignore it and only have access to the sRGB image. The residual image is formed such that it will have negligible pixel values for those portions of the image that lie within the sRGB gamut, and will therefore be highly compressible. Tests on a population of 950 real customer images have demonstrated that the extended dynamic range scene information can be stored with an average file size overhead of about 8% compared to the sRGB images alone. © 2003 Wiley Periodicals, Inc. Col Res Appl, 28, 251–266, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.10160     

Preference for color‐enhanced images assessed by color deficiencies

In this study, we propose color images with color enhancement for congenital red–green color deficiency on the basis of the model proposed by Mochizuki et al. [In Terrassa, Spain: Proceedings for CGIV, Springfield VA: IS&T 2008;208–213] with the aim of facilitating color customization in displays used by color‐deficient observers. Three types of color enhancements for deficiency were addressed in this study: colors were enhanced along the protan confusion line (i.e., P‐enhancement), deutan confusion line (i.e., D‐enhancement), and a line between the protan and deutan confusion lines (i.e., mix‐enhancement). We investigated the color‐enhanced image that is most preferred by each group of protan, deutan, and normal observers. Protan and deutan in this study indicate observers that are both dichromats and anomalous trichromats. Nine protan, nine deutan, and six normal observers participated in the experiment. The results showed that among the three types of color enhancements, the D‐enhancement provided the best performance for protan observers. For deutan observers, all three types provided effective enhancement, although the P‐ and mix‐enhancements were slightly better than the D‐enhancement. Our results indicated that color enhancement increases the preference evaluations of protan and deutan observers. © 2013 Wiley Periodicals, Inc. Col Res Appl, 39, 234–251, 2014; Published Online 20 February 2013 in Wiley Online Library ( wileyonlinelibrary.com ). DOI 10.1002/col.21795     

Determination of oil color by image analysis

An image analysis method was developed to determine oil color based on Lovibond-red according to Chinese standard GB5528-85. For one hue, one feature of image analysis, a negative correlation with the red reading of the Lovibond tintometer (r>0.995) was found. Mathematical models were designed to measure oil color based on image analysis. Fortynine oil samples including peanut oil, soybean oil, fragrant sesame oil, rapeseed oil and cottonseed oil were evaluated to assess the performance of the image analysis models. Good agreement, r ²=0.999, between image analysis values and visual color measurement by using the Lovibond Color Scale was obtained.     

Color errors of digital cameras

The mean color errors of a high‐quality digital camera are defined in CIELAB and CIEDE2000 ΔE units by using 16 ceramic color samples, whose accurate CIELAB values have been measured by a calibrated spectrophotometer. The bandwidths of CCD's color filters are evaluated by taking photographs of CRT‐display primaries. The lowest mean color errors were 13.1 CIELAB ΔE units and 8.1 CIEDE2000 ΔE units before corrections. Large color errors are decreased successfully by using three different methods: simple photoeditor, gamma correction, and multiple regression. © 2004 Wiley Periodicals, Inc. Col Res Appl, 29, 217–221, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.20007     

Segmentation of map image using opponent color dimensions

This article describes a method for segmentation of color geographic map images based on color opponency. In this method, a color-map image is transformed into a color opponent representation as proposed for human vision. The color contrast is enhanced in a manner analogous to the opponent surround receptive fields. The enhancement process separates adjacent foreground and background regions into pairs of opponent colors. The segmentation process can then be easily completed based on this opponent structure. © 1996 John Wiley & Sons, Inc.     

Review and evaluation of color spaces for image/video compression

A color space plays an important role in color image processing and color vision applications. While compressing images/videos, properties of the human visual system are used to remove image details unperceivable by the human eye, appropriately called psychovisual redundancies. This is where the effect of the color spaces' properties on compression efficiency is introduced. In this work, we study the suitability of various color spaces for compression of images and videos. This review work is undertaken in two stages. Initially, a comprehensive review of the published color spaces is done. These color spaces are classified and their advantages, limitations, and applications are also highlighted. Next, the color spaces are quantitatively analyzed and benchmarked in the perspective of image and video compression algorithms, to identify and evaluate crucial color space parameters for image and video compression algorithms.     

Evaluation of a Model of Colour Vision by Magnitude Scalings: Discussion of Collected Results

Coefficients of variation that show the degree to which the Hunt model of colour vision is successful in predicting the LUTCRI appearance data are gathered together in this article for a very wide range of viewing conditions. The model now consistently predicts these data as closely as the results of a single observer. agree with the average results of a group of six observers. The nature of the adjustments that have had to be made to the model to achieve this result are described. and their implications are discussed.     

Visual comfort evaluated by number of categorical colors in a colored image

Visual comfort of a colored image of an ordinary scene was investigated in terms of the number of categorical colors contained in the image. The categorical colors were extracted by presenting the image for 5 s and asking the observer to report all the colors perceived in the image by using categorical color naming. It was found that the number of categorical colors obtained had a high negative correlation with the observer's direct estimation of visual comfort for the image. This indicates that the more colors an image contains, the less comfort for the image is felt, and suggests that visual comfort can be evaluated as a function of the number of categorical colors. The number of categorical colors was also determined colorimetrically and had a high negative correlation to the observer's direct estimation of visual comfort as well. © 2000 John Wiley & Sons, Inc. Col Res Appl, 25, 193–199, 2000     

Let's call it “color‐gamut rendering”

The term “color gamut” historically has been associated with color output such as optimal color stimuli and additive and subtractive imaging systems. Recently, this term has been used with input devices such as scanners and digital cameras. It is proposed that the term “color‐gamut rendering” should be used instead of input devices. This clarifies the distinction between input (analysis) and output (synthesis) color systems in terms of the effect of an input system on defining the colorimetric properties of an output system. © 2007 Wiley Periodicals, Inc. Col Res Appl, 32, 334–335, 2007     

Measurement of memory of color

Colors produced by monochromatic wavelengths of light viewed in isolation have been used as the only visual variables in short‐term delayed matching (DM) and long‐term recall (LTR) protocols to quantify three types of color memory in individuals with normal color vision. Measurements were normally distributed, so that color memories of individuals could be compared in terms of means and standard deviations. The variance of LTR of colors of familiar objects is shown to be separable into two portions, one due to “preferred colors” and the other due to individuals' precisions of matching. The wavelength dependence of DM exhibited minima of standard deviations at the same wavelengths as those reported for color discrimination measured by bipartite wavelength matching, and these wavelengths were shown to occur at the wavelengths of the intersections of cone spectral sensitivities. In an intermediate “green” region of relatively constant color discrimination, it was possible to combine DM measurements for different wavelengths for statistical analysis. The standard deviations of DM for individuals of a healthy population were normally distributed, providing a 95% upper confidence limit for identifying individuals with possible short‐term memory impairment. Preliminary measurements of standard deviations of DM for delay times of ≤ 1 s were consistent with a proposed rapidly decaying color imagery memory. © 2002 Wiley Periodicals, Inc. Col Res Appl, 27, 233–242, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.10067     

A novel Color Rendition Chart for digital tongue image calibration

Digital tongue images are usually acquired by a camera under specific illumination environments. In order to guarantee better color representation of the tongue body, we propose a novel tongue Color Rendition Chart acting as a color reference to be used in color calibration algorithms to standardize the captured tongue images. First, based on a large tongue image database captured with our digital tongue image acquisition system, we establish a statistical tongue color gamut. Then, from the first step, different quantities of colors in the Color Rendition Chart are determined via experimentation. Afterwards, results using X‐Rite's ColorChecker® Color Rendition Chart (a standard in the color calibration field) are compared with the proposed tongue Color Rendition Chart by applying the color difference calculation formula of CIELAB and CIEDE2000 as a reference for the mean color calibration error. The results show that the proposed tongue Color Rendition Chart, which has 24 colors, produces a much smaller error (CIELAB —8.0755/CIEDE 2000—6.3482) compared with X‐Rite's ColorChecker® Color Rendition Chart (CIELAB 1976—14.7836/CIEDE 2000—11.7686). This demonstrates the effectiveness of the novel tongue Color Rendition Chart.     

Applying image‐based color palette for achieving high image quality of displays

In the highly competitive display market, manufacturers continuously develop new technologies to improve the image quality of displays. However, color measurement and visual assessment are time‐consuming to production lines. A new method to measure and improve color quality of the displays automatically therefore, is urgently needed to the manufacturers. This article proposes a familiar color correction strategy to optimize the colors of different displays by means of creating an image‐based color palette which enables color correction for familiar objects (e.g., facial skin, blue sky, or green grass) in the multidisplay systems. To produce the image‐based color palette, the 8‐bit RGB value of each pixel in an image is transformed to L*d*n* (lightness/dominant color/nondominant color) color channels, and the dominant‐color regions in an image are subsequently extracted from the dominant color (d*) channel. The memory color data of familiar objects can be set in reference monitor in advance to determine the dominant color (d*) channel. Then a series of palette colors are generated around a displayed image. The color palette will be displayed as a target for two‐dimensional colorimeter shooting to obtain the measured color data. The familiar color correction model was established based on a first‐order polynomial regression to achieve a polynomial fit between the measured color data and the reference color data on the color palette. The proposed method provides a solution to correct familiar colors on a displayed image, and maintains the original color gamut and tone characteristic in the multidisplay systems simultaneously. It is possible to achieve the preferred intent of the displayed images by using the proposed familiar color correction method. © 2012 Wiley Periodicals, Inc. Col Res Appl, 39, 154–168, 2014     

Heuristic analysis influence of saliency in the color diversity of natural images

The estimation of chromatic diversity of natural images is commonly quantified through the computation of the number of discernible colors and has received much attention because of the different implications it has. However, the relationship between that number and the number of colors that really attracts the attention from an observer is still not clear and has been given little attention. New concepts about salient discernible colors‐the salient chromatic diversity of images‐ and remarkable salient colors‐connected colors in the same salient image area‐are introduced as opposed to the classical number of discernible object colors, which is usually evaluated for the global image without differentiating between probable attended and non‐attended image regions. We have used different well‐known saliency models to locate the salient regions in the scenes and have heuristically studied the extent to which those models preserve the chromatic diversity of natural images. Based on a bottom‐up approach, a reduction of around 40%‐55% in the number of discernible colors were obtained, and not all saliency algorithms preserved a uniform sampling of the original color gamut. Thus, our results suggests that particularly the graph‐based visual saliency model got good low dissimilarity values in comparison with other approaches that put emphasis solely on color as the main low‐level feature. Furthermore, we have introduced a quantification scheme of the average number of remarkable salient colors appearing in the images, and have proved how the heuristic‐based analysis of salient image areas can be used to create segmented images automatically according to their salient chromatic diversity.     

Quadrant dynamic clipped histogram equalization with gamma correction for color image enhancement

The detailed examination of low contrast image is a challenging issue. Thus, it makes difficult for the viewer to bring out the detailed features of the image. Histogram equalization (HE) is an efficient way to intensify the contrast of images. However, classical HE techniques result in immoderate intensification. Hence, an efficient contrast enhancement algorithm called quadrant dynamic clipped HE with gamma correction is proposed. This transformation addresses both over-enhancement and fine detail preservation, which ensures no false contouring. In the proposed method, histogram of the input image is partitioned into four sections using its mean value. Histogram clipping and gamma correction is used to control the color enhancement rate. Then, clipped subhistogram is equalized independently and then they combined together to form an enhanced image. The performance assessment of the proposed and other existing methods is evaluated in terms of entropy, contrast, colorfulness, and saturation. Test results demonstrates that the proposed method outperforms the other existing HE methods in terms of preserving entropy, colorfulness, saturation, and obtaining uniform degree enhancement.     

Transferring the color imagery from an image: A color network model for assisting color combination

In order to transfer the color imagery from an image to another object to generate color schemes, a color network model is proposed. It consists of two subnets: source net for describing the color information of the reference image; target net for demonstrating the target object to be colored. Thus, the problem of reusing features of color patterns is translated into the mapping process from the source net to target net. Four indicators are designed to measure the conformity between the two nets: color sequence, adjacency, concentration, and the subspace size. In the meantime, a comprehensive aesthetic evaluation given by the designer is introduced to search the optimum combination of the four indicators to help generate imagery matching schemes. A prototype system is developed based on CorelDraw as a design tool. The feasibility of the color network model is then verified through a color combination of graphic design task.     

Sensor‐response‐ratio constancy under changes in natural and artificial illuminants

We have analyzed the constancy of the response ratio for cones, second‐stage mechanisms, and CCD sensors when daylight or an artificial illuminant (A, F2, F7, and F11) is changed to an equal‐energy illuminant (E) in scenes containing natural and artificial objects. The response ratios were always found to be roughly constant for all the sensors. For daylight, we have deduced mathematical expressions, which relate the values of these ratios with the correlated color temperature (CCT) and applied these expressions to the synthesis of images of a scene viewed under different daylights corresponding to a variety of CCTs. The results were highly satisfactory in a rural scene for any CCT. In the scene with artificial objects, the results were also good for nonextreme CCTs. We also included in our study artificial illuminants, with which we achieved very good image syntheses for illuminants A and F7. © 2007 Wiley Periodicals, Inc. Col Res Appl, 32, 284–292, 2007     

Color preference of aged observers compared to young observers

In modern society, elderly people tend to become enthusiastic users of displays. These displays are optimized for the visual properties of young adults ignoring the specific attributes of the vision of the elderly though the existence of differences is evident. It is true, however, that most of the investigations follow a bottom‐up paradigm (changes of optical density, neural changes, etc.), and their results are too ambiguous to be taken into account directly in display optimization for the elderly, partly because there is a long‐term adaptation, which apparently recompensates for some of the changes in the human visual system. Preference is a high‐level psychological factor having a very important impact on the acceptability of color displays. The present article follows a top–down methodology to investigate these age‐related differences directly. The first part concentrates on functionalities of the human visual system: white point preference, chroma perception, unique, and preferred hues. Results confirm the long‐term compensation theory for the elderly observers along with the chromatic content decrease of perceived colors. The second part investigates preference differences in case of photorealistic images in terms of global and local contrast, white point, average chroma, and the effect of several image color manipulation techniques. Results indicate significant differences between young and aged observers' color image preference, some of which can be explained with neuro‐physiological changes, others may be attributed to cultural implications. © 2008 Wiley Periodicals, Inc. Col Res Appl, 33, 381–394, 2008     

The feasibility of establishing new color image scales using the magnitude estimation method

Color image is one of the most important factors in art and design. In general, artists and designers apply their own personal image meanings into their work. However, the image meaning for a specific work is frequently in conflict with those of the general observer. Thus it is necessary and important to derive one set of merit color image scales which can be utilized to predict the color image meanings of works in parallel with the average person's perception and which can also serve as a guide for artists and designers. In this study, the psychophysical method (magnitude estimation method), usually used in visual measurement of color appearance, was employed to attempt to establish new color image scales to evaluate the color image meanings of works matching those of the average person. The results show that new color image scales WIP are developed, and the relativity between the latest color image scales WIP and the color attributes (say Lightness L*, Hue h, and Chroma C*) of the CIELAB color space is also discussed. The total mean value of coefficient of variation for the results of visual assessment in the experiment of evaluating the color image meanings of the 207 color specimens used, in general, is about 36, similar to that for those experiments conducted using the psychological method. Also, a good relationship between the new color image scales and the color attributes of the CIELAB color space can be found. © 2007 Wiley Periodicals, Inc. Col Res Appl, 32, 463–468, 2007