Color Research and Application最新文献

Skin color constancy under nonuniform correlated color temperatures (CCT) and multiple light sources has always been a hot issue in color science. A more high-quality skin color reproduction method has broad application prospects in camera photography, face recognition, and other fields. The processing process from the 14bit or 16bit RAW pictures taken by the camera to the final output of 8bit JPG pictures is called the image processing pipeline, in which the steps of the auto-white balance algorithm have a decisive impact on the skin color reproduction result. The traditional automatic white balance algorithm is based on hypothetical statistics. Moreover, the estimated illuminant color is obtained through illuminant estimation. However, the traditional grayscale world, perfect reflector, and other auto-white balance algorithms perform unsatisfactorily under non-uniform or complex light sources. The method based on sample statistics proposes a new solution to this problem from another aspect. The deep learning algorithm, especially the generative adversarial network (GAN) algorithm, is very suitable for establishing the mapping between pictures and has an excellent performance in the fields of image reconstruction, image translation, defogging, and coloring. This paper proposes a new solution to this problem. The asymmetric UNet3+ shape generator integrates better global and local information to obtain a more refined correction matrix incorporating details of the whole image. The discriminator is Patch-discriminator, which focuses more on image details by changing the attention field. The dataset used in this article is the Liverpool-Leeds Skin-color Database (LLSD) and some supplementary images, including the skin color of more than 960 subjects under D65 and different light sources. Finally, we calculate the CIEDE2000 color difference and some other image quality index between the test skin color JPEG picture corrected by the auto-white balance algorithm and the skin color under the corresponding D65 to evaluate the effect of white balance correction. The results show that the asymmetric GAN algorithm proposed in this paper can bring higher quality skin color reproduction results than the traditional auto-white balance algorithm and existing deep learning WB algorithm.

Color palette is a critical component of art, design, and lots of applications, providing the basis for organizing and utilizing colors to achieve specific objectives. However, generating color palettes from digital images presents unique challenges due to the complexity of colors in images. This review comprehensively investigates various techniques for generating color palettes from digital images and provides a thorough classification and discussion of these techniques from multiple perspectives. A color space must be selected to generate a color palette, and a generation method must be employed. This paper offers a concise overview of color spaces, an introduction to current palette generation methods, and an analysis of the metrics used to evaluate color differences between palettes. The review encompasses traditional manual methods and computer-aided automation methods, further categorized as histogram-based, clustering-based, and neural network-based methods. Discussion on the strengths, weaknesses, and applicability of existing methods are presented, and also opportunities for future research to enhance color palette generation from digital images are identified.

Individual differences are a prominent feature of normal color vision and range from variations in sensitivity to perception and color naming. Corrections for differences in spectral sensitivity are common, and there is growing interest in calibrating displays for the sensitivity of an individual observer. In contrast, few studies have explored calibrations for aspects of color appearance. We developed a technique for adjusting images based directly on an individual's hue percepts and illustrate the principle of the approach using a set of hue scaling functions measured previously for a large sample of color-normal observers (Emery et al. PNAS 2023). Colors in an image are mapped onto the average scaling function to define the hue perceived by the average “standard observer.” This hue is then mapped back to the chromaticity that would elicit the same response in any individual observer. With this correction, different observers – each looking at physically different images calibrated for their own hue percepts – should in principle agree on the perceived colors. Adjustments of this kind could be easily implemented on standard displays, because they require only measures of hue percepts and not spectral sensitivity, and could lead to greater consistency in the perception and communication about color across individuals with potentially widely different perceptual experiences of color.

Garden color plays a key role in creating a spatial atmosphere, artistic conception and regional landscape characteristics. The Chinese classical garden has a profound traditional Chinese culture at its core, and its color collocation is a microcosm of Chinese culture. An in-depth exploration of the color matching of classical gardens is helpful for promoting the organic combination of modern urban construction needs and traditional culture. Therefore, this paper takes Tang-style gardens in Xi'an as an example and selects five typical classical gardens to study the color characteristics of garden elements, uses natural color system (NCS) to record color data from classical garden elements, establishes a color map, and quantitatively evaluates the color harmony of classical garden elements on the basis of Moon Spencer (M-S) color harmony theory. Research has revealed that the color system of Tang-style gardens in Xi'an is unique and includes 244 NCS standard colors with a wide distribution of hues. The study distinguishes between humanistic environment colors and natural environment colors that vary greatly with season, and provides a wide color range that reflects the beauty of nature. The color range of the humanistic environment appears calm and solemn, with colors having low blackness and low chromaticness in general. There is a certain difference in color harmony between natural and cultural environment colors, but the overall color harmony is good. This paper not only enriches the color theory system of Tang-style gardens but also provides a reference for promoting the scientific application of traditional urban colors in modern urban landscape construction.

Just-noticeable color difference (JNCD) is important in color specification and characterization. The commonly referenced specification of JNCD (i.e., 0.004 or 0.0033 u′v′ unit) is thought to originate from the MacAdam ellipses, which were derived using 2° color stimuli and characterized using the CIE 1931 2° color matching functions (CMFs). However, there is no universally agreed or clear definition of JNCD. Also, such a specification is widely used in various ways based on an assumption that it is applicable regardless of the actual size of the stimuli and CMFs. In this study, an experiment using a constant stimuli method was carried out. The human observers evaluated a series pairs of test and reference stimuli, with a field of view (FOV) of 2° or 10°. The chromaticities of the test stimuli were carefully calibrated using four standard CIE CMFs (i.e., CIE 1931 2°, 1964 10°, 2006 2°, and 10° CMFs). The results suggested that the widely used specification of JNCD seems to be derived based on the one standard deviation ellipses, the use of these four CMFs has little effect on the specification, and the JNCD value for stimuli with an FOV of 10° is 0.0025–0.0027 u′v′ unit depending on the CMFs.

In 2022, CIE recommend the CIECAM16 color appearance model to replace the current CIECAM02 model for color management systems. In this paper we will report on how CIECAM16 was developed, describe the differences between CIECAM16 and CIECAM02, describe the phenomena CIECAM16 can predict, and show the performance of CIECAM16 in the prediction of perceptual color attributes. We will then discuss the numerical determination of the domain and range of the CIECAM16 forward transformation. Finally, the domain and ranges in various conditions are visualized. CIECAM16 is capable of the accurate prediction of color appearance under a wide range of viewing conditions and with the domain and range problems now solved, it can be effectively applied to cross-media color image reproduction. It can also be used to estimate the color rendering properties of light sources and for the establishment of a uniform color space for color difference evaluation.