Efficient quantization of painting images by relevant colors.

Realistic images often contain complex variations in color, which can make economical descriptions difficult. Yet human observers can readily reduce the number of colors in paintings to a small proportion they judge as relevant. These relevant colors provide a way to simplify images by effectively quantizing them. The aim here was to estimate the information captured by this process and to compare it with algorithmic estimates of the maximum information possible by colorimetric and general optimization methods. The images tested were of 20 conventionally representational paintings. Information was quantified by Shannon's mutual information. It was found that the estimated mutual information in observers' choices reached about 90% of the algorithmic maxima. For comparison, JPEG compression delivered somewhat less. Observers seem to be efficient at effectively quantizing colored images, an ability that may have applications in the real world.

Band Selection for Dehazing Algorithms Applied to Hyperspectral Images in the Visible Range.

Images captured under bad weather conditions (e.g., fog, haze, mist, dust, etc.), suffer from poor contrast and visibility, and color distortions. The severity of this degradation depends on the distance, the density of the atmospheric particles and the wavelength. We analyzed eight single image dehazing algorithms representative of different strategies and originally developed for RGB images, over a database of hazy spectral images in the visible range. We carried out a brute force search to find the optimum three wavelengths according to a new combined image quality metric. The optimal triplet of monochromatic bands depends on the dehazing algorithm used and, in most cases, the different bands are quite close to each other. According to our proposed combined metric, the best method is the artificial multiple exposure image fusion (AMEF). If all wavelengths within the range 450-720 nm are used to build a sRGB renderization of the imagaes, the two best-performing methods are AMEF and the contrast limited adaptive histogram equalization (CLAHE), with very similar quality of the dehazed images. Our results show that the performance of the algorithms critically depends on the signal balance and the information present in the three channels of the input image. The capture time can be considerably shortened, and the capture device simplified by using a triplet of bands instead of the full wavelength range for dehazing purposes, although the selection of the bands must be performed specifically for a given algorithm.

Psychophysical Determination of the Relevant Colours That Describe the Colour Palette of Paintings.

In an early study, the so-called "relevant colour" in a painting was heuristically introduced as a term to describe the number of colours that would stand out for an observer when just glancing at a painting. The purpose of this study is to analyse how observers determine the relevant colours by describing observers' subjective impressions of the most representative colours in paintings and to provide a psychophysical backing for a related computational model we proposed in a previous work. This subjective impression is elicited by an efficient and optimal processing of the most representative colour instances in painting images. Our results suggest an average number of 21 subjective colours. This number is in close agreement with the computational number of relevant colours previously obtained and allows a reliable segmentation of colour images using a small number of colours without introducing any colour categorization. In addition, our results are in good agreement with the directions of colour preferences derived from an independent component analysis. We show that independent component analysis of the painting images yields directions of colour preference aligned with the relevant colours of these images. Following on from this analysis, the results suggest that hue colour components are efficiently distributed throughout a discrete number of directions and could be relevant instances to a priori describe the most representative colours that make up the colour palette of paintings.

Eight-Channel Multispectral Image Database for Saliency Prediction.

Saliency prediction is a very important and challenging task within the computer vision community. Many models exist that try to predict the salient regions on a scene from its RGB image values. Several new models are developed, and spectral imaging techniques may potentially overcome the limitations found when using RGB images. However, the experimental study of such models based on spectral images is difficult because of the lack of available data to work with. This article presents the first eight-channel multispectral image database of outdoor urban scenes together with their gaze data recorded using an eyetracker over several observers performing different visualization tasks. Besides, the information from this database is used to study whether the complexity of the images has an impact on the saliency maps retrieved from the observers. Results show that more complex images do not correlate with higher differences in the saliency maps obtained.

How daylight influences high-order chromatic descriptors in natural images.

Despite the global and local daylight changes naturally occurring in natural scenes, the human visual system usually adapts quite well to those changes, developing a stable color perception. Nevertheless, the influence of daylight in modeling natural image statistics is not fully understood and has received little attention. The aim of this work was to analyze the influence of daylight changes in different high-order chromatic descriptors (i.e., color volume, color gamut, and number of discernible colors) derived from 350 color images, which were rendered under 108 natural illuminants with Correlated Color Temperatures (CCT) from 2735 to 25,889 K. Results suggest that chromatic and luminance information is almost constant and does not depend on the CCT of the illuminant for values above 14,000 K. Nevertheless, differences between the red-green and blue-yellow image components were found below that CCT, with most of the statistical descriptors analyzed showing local extremes in the range 2950 K-6300 K. Uniform regions and areas of the images attracting observers' attention were also considered in this analysis and were characterized by their patchiness index and their saliency maps. Meanwhile, the results of the patchiness index do not show a clear dependence on CCT, and it is remarkable that a significant reduction in the number of discernible colors (58% on average) was found when the images were masked with their corresponding saliency maps. Our results suggest that chromatic diversity, as defined in terms of the discernible colors, can be strongly reduced when an observer scans a natural scene. These findings support the idea that a reduction in the number of discernible colors will guide visual saliency and attention. Whatever the modeling is mediating the neural representation of natural images, natural image statistics, it is clear that natural image statistics should take into account those local maxima and minima depending on the daylight illumination and the reduction of the number of discernible colors when salient regions are considered.

Colorimetric analysis of outdoor illumination across varieties of atmospheric conditions.

Solar illumination at ground level is subject to a good deal of change in spectral and colorimetric properties. With an aim of understanding the influence of atmospheric components and phases of daylight on colorimetric specifications of downward radiation, more than 5,600,000 spectral irradiance functions of daylight, sunlight, and skylight were simulated by the radiative transfer code, SBDART [Bull. Am. Meteorol. Soc.79, 2101 (1998)], under the atmospheric conditions of clear sky without aerosol particles, clear sky with aerosol particles, and overcast sky. The interquartile range of the correlated color temperatures (CCT) for daylight indicated values from 5712 to 7757 K among the three atmospheric conditions. A minimum CCT of ∼3600 K was found for daylight when aerosol particles are present in the atmosphere. Our analysis indicated that hemispheric daylight with CCT less than 3600 K may be observed in rare conditions in which the level of aerosol is high in the atmosphere. In an atmosphere with aerosol particles, we also found that the chromaticity of daylight may shift along the green-purple direction of the Planckian locus, with a magnitude depending on the spectral extinction by aerosol particles and the amount of water vapor in the atmosphere. The data analysis showed that an extremely high value of CCT, in an atmosphere without aerosol particles, for daylight and skylight at low sun, is mainly due to the effect of Chappuis absorption band of ozone at ∼600 nm. In this paper, we compare our data with well-known observations from previous research, including the ones used by the CIE to define natural daylight illuminants.

Nonrigid registration with free-form deformation model of multilevel uniform cubic B-splines: application to image registration and distortion correction of spectral image cubes.

In spectral imaging, spatial and spectral information of an image scene are combined. There exist several technologies that allow the acquisition of this kind of data. Depending on the optical components used in the spectral imaging systems, misalignment between image channels can occur. Further, the projection of some systems deviates from that of a perfect optical lens system enough that a distortion of scene content in the images becomes apparent to the observer. Correcting distortion and misalignment can be complicated for spectral image data if they are different at each image channel. In this work, we propose an image registration and distortion correction scheme for spectral image cubes that is based on a free-form deformation model of uniform cubic B-splines with multilevel grid refinement. This scheme is adaptive with respect to image size, degree of misalignment, and degree of distortion, and in that sense is superior to previous approaches. We support our proposed scheme with empirical data from a Bragg-grating-based hyperspectral imager, for which a registration accuracy of approximately one pixel was achieved.

Generalized inverse-approach model for spectral-signal recovery.

We have studied the transformation system of a spectral signal to the response of the system as a linear mapping from higher to lower dimensional space in order to look more closely at inverse-approach models. The problem of spectral-signal recovery from the response of a transformation system is generally stated on the basis of the generalized inverse-approach theorem, which provides a modular model for generating a spectral signal from a given response value. The controlling criteria, including the robustness of the inverse model to perturbations of the response caused by noise, and the condition number for matrix inversion, are proposed, together with the mean square error, so as to create an efficient model for spectral-signal recovery. The spectral-reflectance recovery and color correction of natural surface color are numerically investigated to appraise different illuminant-observer transformation matrices based on the proposed controlling criteria both in the absence and the presence of noise.

Spectral recovery of outdoor illumination by an extension of the Bayesian inverse approach to the Gaussian mixture model.

The Bayesian inference approach to the inverse problem of spectral signal recovery has been extended to mixtures of Gaussian probability distributions of a training dataset in order to increase the efficiency of estimating the spectral signal from the response of a transformation system. Bayesian (BIC) and Akaike (AIC) information criteria were assessed in order to provide the Gaussian mixture model (GMM) with the optimum number of clusters within the spectral space. The spectra of 2600 solar illuminations measured in Granada (Spain) were recovered over the range of 360-830 nm from their corresponding tristimulus values using a linear model of basis functions, the Wiener inverse (WI) method, and the Bayesian inverse approach extended to the GMM (BGMM). A model of Gaussian mixtures for solar irradiance was deemed to be more appropriate than a single Gaussian distribution for representing the probability distribution of the solar spectral data. The results showed that the estimation performance of the BGMM method was better than either the linear model or the WI method for the spectral approximation of daylight from the three-dimensional tristimulus values.

Contrast edge colors under different natural illuminations.

Essential to sensory processing in the human visual system is natural illumination, which can vary considerably not only across space but also along the day depending on the atmospheric conditions and the sun's position in the sky. In this work, edges derived from the three postreceptoral Luminance, Red-Green, and Blue-Yellow signals were computed from hyperspectral images of natural scenes rendered with daylights of Correlated Color Temperatures (CCTs) from 2735 to 25,889 K; for low CCT, the same analysis was performed using Planckian illuminants up to 800 K. It was found that average luminance and chromatic edge contrasts were maximal for low correlated color temperatures and almost constants above 10,000 K. The magnitude of these contrast changes was, however, only about 2% across the tested daylights. Results suggest that the postreceptoral opponent and nonopponent color vision mechanisms produce almost constant responses for color edge detection under natural illumination.

Trichromatic red-green-blue camera used for the recovery of albedo and reflectance of rough-textured surfaces under different illumination conditions.

Photometric-stereo techniques are based on the fact that image intensity depends upon the orientation of the surface with regard to the source of the illumination and its spectral reflectance. They are of special interest when dealing with rough surfaces because they usually present shadowed regions where sudden illumination changes might be found. In the present work we introduce an extension of the four-source photometric-stereo algorithm to color images that is able to recover the surface spectral reflectance of objects captured with a red-green-blue (RGB) camera. This method allows image rendering, even for rough-textured surfaces, under different directions of the impinging illumination. In addition, the introduction of spectral recovery techniques applied to the albedo and spectral reflectance from rough surfaces offers the possibility of image rendering for scenes captured under sources of illumination differing in spectral distribution. Using albedo instead of RGB information helps to avoid any shadows or highlights that might falsify results. One of the advantages of this spectral-based photometric-stereo method is that it can recover not only the albedo values, but also the spectral reflectance spectrum of an object's surface on a pixel-by-pixel basis, as can be done with more complex hyperspectral imaging devices involving a camera coupled to an extensive set of narrowband filters.