Recovery of reflectance spectra from CIE tristimulus values using a progressive database selection technique

Abstract

A procedure for creating efficient reflectance spectra from CIE tristimulus colour values is described using a modified linear model. By fixing certain criteria based on colour difference values, the proposed technique preliminarily selects a series of suitable samples from a main data set containing the reflectance values of a large number of different coloured samples, based on the colour specifications of a given sample. In this way, a series of different databases containing the reflectance values of confirmed samples relating to the particular samples are formed. Then, a well‐known principal components linear model is applied using three basis functions. This operation yields higher cumulative variances in comparison with the original database, having the same number of basis vectors. The performance of the proposed method is evaluated using a different collection of samples and the resulting spectra show considerable improvements in terms of root mean square error as well as colour difference values under different illuminants.

Optimized basis function for spectral reflectance recovery from tristimulus values

“…[17][18][19] A multispectral camera system allows the reflectance spectra to be recovered more accurately than a trichromatic camera system. However, such multispectral systems are expensive and the recovery procedure is complex.…”

“…[17][18][19] It applies singular value decomposition (SVD) for a set of reflectance spectra data. By this method, several basis functions are obtained and used to create a low-dimensional representation for approximating the reflectance spectra.…”

“…Garcia et al 7) proposed to cluster samples' hues into subgroups to improve the accuracy of the basis functions in the process of calculation by the PCA method. Ansari et al 17) presented a PCA recovery method in which a series of samples is selected from a large number of different reflectance spectra data. Ayala et al 18) improved the accuracy by dividing reflectance spectra data into ten subgroups based on the hue similarity of samples.…”

“…However, in the first kind of method 17) the reflectance spectra data set were divided into several subgroups. In such methods, different basis functions are calculated for different subgroups by the PCA method.…”

Measuring Reflectance Spectra of Textile Fabrics by Scanner

“…By grading the eigenvectors for descending eigenvalues, so that largest is first, one can create an ordered orthogonal method with the first eigenvector having the direction of largest variance of the data. In this way, we can find directions in which the data set has the most significant amounts of energy and variation [13][14][15][16][17][18][19].…”