基于混合色适应转换的图像色差模型研究

由于不同的图像输出媒介有着不同特性,评价不同输出媒介的输出图像质量比较困难。以常见的软拷贝输出及硬拷贝输出为研究对象,建立了基于混合色适应转换的图像色差模型,同时设计了相应的心理物理学实验来验证模型计算结果。研究表明,基于混合色适应转换的图像色差模型,在计算不同媒介输出图像色差时表现良好,基本可以作为评价软拷贝输出与硬拷贝输出图像色差的工具。     

Ratio model for suprathreshold hue-increment detection.

We use psychophysical techniques to investigate the neural mechanisms subserving suprathreshold chromatic discrimination in human vision. We address two questions: (1) How are the postreceptoral detection mechanism responses combined to form suprathreshold chromatic discriminators? and (2) How do these discriminators contribute to color perception? We use a pedestal paradigm in which the subject is required to distinguish between a pedestal stimulus and the same pedestal added to a chromatic increment (the test). Our stimuli are represented in a cardinal space, in which the axes express the responses of the three postreceptoral detection mechanisms normalized relative to their respective detection thresholds. In the main experiment the test (a hue increment) was fixed in the direction orthogonal to the pedestal in our cardinal space. We found that, for high pedestal contrasts, the test threshold varied proportionally with the pedestal contrast. This result suggests the presence of a hue-increment detector dependent on the ratio of the outputs from the red-green and blue-yellow postreceptoral detection mechanisms. The exception to this was for pedestals and tests fixed along the cardinal axes. In that case detection was enhanced by direct input from the postreceptoral mechanism capable of detecting the test in isolation. Our results also indicate that discrimination in the red-green/luminance and blue-yellow/luminance planes exhibits a behavior similar to discrimination within the isoluminant plane. In the final experiment we observed that thresholds for hue-increment identification (e.g., selecting the bluer of two stimuli) are also governed by a ratio relationship. This finding suggests that our ratio-based mechanisms play an important role in color-difference perception.     

Opponent chromatic mechanisms: relation to photopigments and hue naming.

Opponent chromatic response functions were determined from monochromatic, equal-luminance stimuli from 400 to 700 nm for three observers using a hue cancellation procedure. The same observers scaled the hue of the stimuli using the terms red, green, yellow, and blue. The results showed that the hue scaling was accurately predicted from the cancellation functions using the model of Hurvich and Jameson. Theoretical curves were generated to fit the chromatic response functions with a linear combination of three cone photopigments. The theoretical photopigments were based on an idopsin nomogram with lambdamax at a = 435, beta = 530, and lambda = 562 nm. An estimate of the density of each observer's preretinal optic media was obtained in order to relate the photopigment absorption spectra to the psychophysical data. Good linear fits were obtained for each observer's red-green curve, but not for the yellow-blue curves. A nonlinear model with an expansive exponent was used to fit the yellow-blue response functions with the three theoretical photopigments.     

Effective contrast of colored stimuli in the mesopic range: a metric for perceived contrast based on achromatic luminance contrast

Little is known about how color signals and cone- and rod-based luminance signals contribute to perceived contrast in the mesopic range. In this study the perceived contrast of colored, mesopic stimuli was matched with that of spatially equivalent achromatic stimuli. The objective was to develop a metric for perceived contrast in the mesopic range in terms of an equivalent achromatic luminance contrast, referred to here as effective contrast. Stimulus photopic luminance contrast, scotopic luminance contrast, and chromatic difference from the background all contributed to effective contrast over the mid-mesopic range, but their contributions were not independent and varied markedly with background luminance. Surprisingly, color made a significant contribution to effective contrast from 10 to approximately 0.003 cd m(-2). A model describing this relationship is introduced (R2 = 0.89) and compared with predictions of mesopic luminance contrast obtained from a number of models proposed as systems of mesopic photometry.     

Calculation of retinal image quality for polychromatic light

Although the retinal image is typically polychromatic, few studies have examined polychromatic image quality in the human eye. We begin with a conceptual framework including the formulation of a psychophysical linking hypothesis that underlies the utility of image quality metrics based on the polychromatic point-spread function. We then outline strategies for computing polychromatic point-spread functions of the eye when monochromatic aberrations are known for only a single wavelength. Implementation problems and solutions for this strategy are described. Polychromatic image quality is largely unaffected by wavelength-dependent diffraction and higher-order chromatic aberration. However, accuracy is found to depend critically upon spectral sampling. Using typical aberrations from the Indiana Aberration Study, we assessed through-focus image quality for model eyes with and without chromatic aberrations using a polychromatic metric called the visual Strehl ratio. In the presence of typical levels of monochromatic aberrations, the effect of longitudinal chromatic aberration is greatly reduced. The effect of typical levels of transverse chromatic aberration is virtually eliminated in the presence of longitudinal chromatic aberration and monochromatic aberrations. Clinical value and limitations of the method are discussed.     

Evaluation of single-pigment shift model of anomalous trichromacy.

The spectral sensitivity of the visual photopigments, the interobserver variability in color judgments, and the spectral locus of unique yellow provide three major problems for accounts of X-chromosomal-linked anomalous trichromacy. According to the single-pigment hypothesis, the primary defect in anomalous trichromacy is a wavelength shift in the peak sensitivity of one of the three visual photopigments. We show that this shift results in reduction of the anomalous trichromat's r-g opponent chromatic channel. The distribution of response variability in Rayleigh equation match widths due to factors other than the spectral characteristics of the photopigments is similar in normal and anomalous trichromats. When normal and anomalous trichromats make hue estimations of sets of stimuli designed to contain similar chromatic information, their judgments show similar variability. Calculation of the r-g opponent chromatic channel can provide correct predictions of the spectral loci for unique yellow for anomalous trichromats.     

基于iCAM的印刷品在线颜色复制质量的检测与评价

目的为了提高印刷品在线检测中颜色复制质量评价的准确性,构建基于iCAM图貌模型的印刷品在线检测色差计算模型。方法将印刷品在线检测系统中CCD相机拍摄的待检测样张按照墨区进行分区,基于iCAM图貌模型和IPT均匀颜色空间建立在线色差检测模型,并计算样张各个墨区的色差;设计视觉心理学实验评价样张各墨区颜色的复制质量;采用STRESS值评价基于iCAM的色差检测模型计算出的色差与视觉感受色差的一致性,并与基于CIELAB的色差进行比较,验证该模型的在线色差预测精度。结果基于iCAM的色差模型计算结果具有较高的视觉一致性,更符合视觉评价结果,且与传统色差模型相比,具有更高的在线色差检测精度。结论基于iCAM的印刷品在线检测色差计算模型考虑了人眼视觉的空间特性及色适应,适用于复杂观察条件,能提高印刷品在线检测中的色差检测精度,可用于实际生产中印刷品在线颜色复制质量的检测与评价。     

L/M cone ratios in human trichromats assessed by psychophysics, electroretinography, and retinal densitometry

Estimates of the relative numbers of long-wavelength-sensitive (L) and middle-wavelength-sensitive (M) cones vary considerably among normal trichromats and depend significantly on the nature of the experimental method employed. Here we estimate L/M cone ratios in a population of normal observers, using three psychophysical tasks-detection thresholds for cone-isolating stimuli at different temporal frequencies, heterochromatic flicker photometry, and cone contrast ratios at minimal flicker perception--as well as flicker electroretinography and retinal densitometry. The psychophysical tasks involving high temporal frequencies, specifically designed to tap into the luminance channel, provide average L/M cone ratios that significantly differ from unity with large interindividual variation. In contrast, the psychophysical tasks involving low temporal frequencies, chosen to tap into the red-green chromatic channel, provide L/M cone ratios that are always close to unity. L/M cone ratios determined from electroretinographic recordings or from retinal densitometry correlate with those determined from the high-temporal-frequency tasks. These findings suggest that the sensitivity of the luminance channel is directly related to the relative densities of the L and the M cones and that the red-green chromatic channel introduces a gain adjustment to compensate for differences in L and M cone signal strength.     

Differential equation for normal glass dispersion and evaluation of the secondary spectrum

Some of the questions concerning secondary chromatic aberration at both sides of the visible band of the spectrum are the following: (1) What is the bandwidth at different wavelengths, given the permissible chromatic aberration circle and the lens aperture? (2) What is the size of the chromatic aberration circle, given the wavelength, the bandwidth, and the lens aperture? The answers to these and other questions may be found with the new definitions of V-number and relative partial dispersion P based on infinitesimal bandwidths that we propose. In addition, an alignment chart for the secondary color of a normal glass doublet is presented, so fast answers to the questions posed above and to other questions concerned with secondary color can be found. In addition, a continual challenge in computer-aided lens design is the use of optical glasses as design parameters in simultaneous optimization of lens systems over various regions of the spectrum. This problem could be solved if we could find an ideal glass family, not too different from real glasses, such that, given the refractive index n and the V-number at any wavelength, the indices at all wavelengths could be determined. Therefore we derive a differential equation for normal glass dispersion and present a recursive solution.     

The Verriest Lecture: color lessons from space, time and motion

The appearance of a chromatic stimulus depends on more than the wavelengths composing it. The scientific literature has countless examples showing that spatial and temporal features of light influence the colors we see. Studying chromatic stimuli that vary over space, time, or direction of motion has a further benefit beyond predicting color appearance: the unveiling of otherwise concealed neural processes of color vision. Spatial or temporal stimulus variation uncovers multiple mechanisms of brightness and color perception at distinct levels of the visual pathway. Spatial variation in chromaticity and luminance can change perceived three-dimensional shape, an example of chromatic signals that affect a percept other than color. Chromatic objects in motion expose the surprisingly weak link between the chromaticity of objects and their physical direction of motion, and the role of color in inducing an illusory motion direction. Space, time, and motion-color's colleagues-reveal the richness of chromatic neural processing.     

Parametric fuzzy sets for automatic color naming

In this paper we present a parametric model for automatic color naming where each color category is modeled as a fuzzy set with a parametric membership function. The parameters of the functions are estimated in a fitting process using data derived from psychophysical experiments. The name assignments obtained by the model agree with previous psychophysical experiments, and therefore the high-level color-naming information provided can be useful for different computer vision applications where the use of a parametric model will introduce interesting advantages in terms of implementation costs, data representation, model analysis, and model updating.     

Recognition of polychromatic three-dimensional objects

We propose to use optical multichannel correlation in various chromatic systems to obtain a setup for recognition of polychromatic three-dimensional (3-D) objects based on Fourier-transform profilometry. Because red-green-blue color components are not able to split the luminance information of objects in a defined component, when the 3-D objects are brighter than the reference objects the correlation result gives false alarms. We demonstrate that it is possible to use different color spaces that can split luminance from chromatic information to yield adequate recognition of polychromatic 3-D objects. We show experimental results that prove the utility of the proposed method.     

Theory of spatiochromatic image encoding and feature extraction

We consider how to interpret, filter, and cross-correlate complex-value color (hue and saturation) images by using a single discrete Fourier transform: the spatiochromatic discrete Fourier transform. The model defines new types of spatiochromatic oriented sinusoidal gratings, termed rainbow gratings, which encode the variation of color over space. We demonstrate how color-opponent detectors observed within the vertebrate visual system can be easily defined by linear filters within this representation. This model also allows us to filter and detect both spatial and chromatic patterns in images by using a single cross-correlation procedure. In doing so, we explore a new form of the Cauchy-Schwartz inequality applied to complex-valued scalar products. Results demonstrate the power of this form of spatiochromatic matched filtering in detecting signals embedded in such a significant amount of noise that they are not visible to the unaided human eye.     

1/f Noise in human color vision: the role of S-cone signals

We examine the functional role of S-cone signals on reaction time (RT) variability in human color vision. Stimuli were selected along red-green and blue-yellow cardinal directions and at random directions in the isoluminant plane of the color space. Trial-to-trial RT variability was not statistically independent but correlated across experimental conditions and exhibited 1/f noise spectra with an exponent close to unity in most of the cases. Regarding contrast coding, 1/f noise for random chromatic stimuli at isoluminance was similar to that for achromatic stimuli, thus suggesting that S-cone signals reduce variability of higher order color mechanisms. If we regard spatial coding, the effect of S-cone density in the retina on RT variability was investigated. The magnitude of 1/f noise at 16 min of arc (S-cone free zone) was higher than at 90 min of arc in the blue-yellow channel, and it was similar for the red-green channel. The results suggest that S-cone signals are beneficial and they modulate 1/f noise spectra at postreceptoral stages. The implications related to random multiplicative processes as a possible source of 1/f noise and the optimal information processing in color vision are discussed.     

Predicting color from gray: the relationship between achromatic adjustment and asymmetric matching.

Achromatic adjustment has been used widely to study color context effects. In the achromatic adjustment procedure, an observer adjusts a test stimulus until it appears black, gray, or white. By its nature, achromatic adjustment directly measures the effect of context only for stimuli that appear gray. We present achromatic loci measured in two contexts and asymmetric color matches measured across the same two contexts. The results indicate that achromatic adjustments, together with a gain-control model, may be used to make accurate predictions of the chromaticity of asymmetric matches. Thus measurements of the effect of context for test stimuli that appear gray may be used to predict the effect of context for stimuli that appear colored. The experiments also indicate that accurate prediction depends on ensuring that observers use similar fixational strategies for the two judgments.     

Highly flexible transparent film heaters based on random networks of silver nanowires

We demonstrate a new concept for the fabrication of flexible transparent thin film heaters based on silver nanowires. Thanks to the intrinsic properties of random networks of metallic nanowires, it is possible to combine bendability, transparency and high heating performances at low voltage, typically below 12 V which is of interest for many applications. This is currently not possible with transparent conductive oxide technologies, and it compares well with similar devices fabricated with carbon nanotubes or graphene. We present experiments on glass and poly(ethylene naphthalate) (PEN) substrates (with thicknesses of 125 μm and extremely thin 1.3 μm) with excellent heating performances. We point out that the amount of silver necessary to realize the transparent heaters is very low and we also present preliminary results showing that this material can be efficiently used to fabricate photochromic displays. To our knowledge, this is the first report of metallic nanowire-based transparent thin film heaters. We think these results could be a useful approach for the engineering of highly flexible and transparent heaters which are not attainable by existing processes.      

The relationship between peripherally matched invariant hues and unique hues: a cone-contrast approach

A characteristic shift in hue and saturation occurs when colored targets are viewed peripherally compared with centrally. Four hues, one in each of the red, blue, green, and yellow regions of color space, remain unchanged when presented in the peripheral field. Apart from green, these peripherally invariant hues correspond almost exactly in color space with the unique hues. We explore this puzzling observation using asymmetric color-matching and color-naming experiments and computing cone contrasts for peripheral and central stimuli. We find that the difference between cone contrasts for the peripheral and central stimuli reaches a maximum at the chromatic axis corresponding to peripherally invariant green. We speculate that the effect is linked to a weakened signal from M-cones and probably associated with a reduced number of M-cones in peripheral retina.     

真彩色动感全息图

提出了用单色激光器制作真彩色动感全息图的方法。该方法用三色光栅相机将景物的彩色信息记录在一张透明片上,并使用散射参光法一次曝光合成一幅真彩色全息图。对从不同观察角度所拍摄的透明片进行多次全息曝光并记录在同一张全息片上,可在普通白炽灯下看到景物的真彩色动感全息像。采用散射光作为参考光记录全息片,大大降低了相干噪声。一次曝光记录一幅真彩色全息图,不仅使透明片数量和曝光次数减少为分色记录的1／3,而且消除了由分色记录三次曝光所带来的颜色分离,增加了彩色像的真实感和清晰度。     

Variations in normal color vision. II. Unique hues

We examined individual differences in the color appearance of nonspectral lights and asked how they might be related to individual differences in sensitivity to chromatic stimuli. Observers set unique hues for moderately saturated equiluminant stimuli by varying their hue angle within a plane defined by the LvsM and SvsLM cone-opponent axes that are thought to characterize early postreceptoral color coding. Unique red settings were close to the +L pole of the LvsM axis, while green, blue, and yellow settings clustered along directions intermediate to the LvsM and SvsLM axes and thus corresponded to particular ratios of LvsM to SvsLM activity. Interobserver differences in the unique hues were substantial. However, no relationship was found between hue settings and relative sensitivity to the LvsM and SvsLM axes. Moreover, interobserver variations in different unique hues were uncorrelated and were thus inconsistent with a common underlying factor such as relative sensitivity or changes in the spectral sensitivities of the cones. Thus for the moderately saturated lights we tested, the unique hues appear largely unconstrained by normal individual differences in the cone-opponent axes. In turn, this suggests that the perceived hue for these stimuli does not depend on fixed (common) physiological weightings of the cone-opponent axes or on fixed (common) color signals in the environment.