Daylight illuminant retrieval using redundant image elements

We present a method for retrieving illuminant spectra from a set of images taken with a fixed location camera, such as a surveillance or panoramic one. In these images, there will be significant changes in lighting conditions and scene content, but there will also be static elements in the background. As color constancy is an under-determined problem, we propose to exploit the redundancy and constancy offered by the static image elements to reduce the dimensionality of the problem. Specifically, we assume that the reflectance properties of these objects remain constant across the images taken with a given fixed camera. We demonstrate that we can retrieve illuminant and reflectance spectra in this framework by modeling the redundant image elements as a set of synthetic RGB patches. We define an error function that takes the RGB patches and a set of test illuminants as input and returns a similarity measure of the redundant surfaces reflectances. The test illuminants are then varied until the error function is minimized, returning the illuminants under which each image in the set was captured. This is achieved by gradient descent, providing an optimization method that is robust to shot noise.

Exploiting redundancy in color images

Joanna Marguier

The colors in a digital image do not only depend on the scene's reflectances, but also on the lighting conditions and on the characteristics of the imaging device. It is often necessary to separate the influence of these three elements, in particular to separate the contribution of the object's reflection properties from the one of the incident light. Both are usually represented by spectral quantities, i.e., defined over the visible spectrum. The retrieval of illuminant and reflectances from an RGB image, which provides three values per pixel, is an under-determined problem and requires additional information or assumption on the scene's content in order to be solved. In this thesis, we study how redundant information across a set of images can be exploited to recover either illuminant or reflectance knowledge. We demonstrate that accurate object color can be retrieved as tristimulus values from standard RGB images captured with an unknown camera under uncontrolled lighting conditions and no access to raw data, but containing a limited number of reference colors in the form of a specific calibration target. The correction must be constrained to a limited range of colors similar to the one of the object of interest, which are compared to pre-computed references for the derivation of a linear transform. We implement this method for two consumer oriented applications in Makeup and Home Décor, respectively, which provide users with color advice. The correction accuracy is of ΔE*ab ≃ 1 to 2 depending on the design of the target. We show that this accuracy is sufficient for the applications at hand. We also developed a method to retrieve illuminant spectra from a set of images taken with fixed location cameras, such as panoramic or surveillance ones. In pictures captured with such devices, there will be changes in lighting and dynamic content, but there will also be constant objects. We propose to use these elements as reference colors to solve for color constancy. More precisely, we exploit that their reflectances, while unknown, remain constant across images. We invert a series of image formation models in parallel for a set of test illuminants and, by forcing the output reflectances to match, deduce the illuminant under which each image was captured. Simulations on synthetic RGB patches demonstrate that real daylight illuminants can be retrieved with a median angular under 3° from sets containing as little as four images using a limited number of reference surfaces.

EPFL2010

Illuminant retrieval for fixed location cameras

Joanna Marguier, Sabine Süsstrunk

Fixed location cameras, such as panoramic cameras or surveillance cameras, are very common. In images taken with these cameras, there will be changes in lighting and dynamic image content, but there will also be constant objects in the background. We propose to solve for color constancy in this framework. We use a set of images to recover the scenes’ illuminants using only a few surfaces present in the scene. Our method retrieves the illuminant in every image by minimizing the difference between the reflectance spectra of the redundant elements’ surfaces or, more precisely, between their corresponding sensor response values. It is assumed that these spectra are constant across images taken under different illuminants. We also recover an estimate of the reflectance spectra of the selected elements. Experiments on synthetic and real images validate our method.

2008

Exploiting redundancy in color images

Joanna Marguier

EPFL2010