Regularized learning framework in the estimation of reflectance spectra from camera responses

Abstract

For digital cameras, device-dependent pixel values describe the camera's response to the incoming spectrum of light. We convert device-dependent RGB values to device- and illuminant-independent reflectance spectra. Simple regularization methods with widely used polynomial modeling provide an efficient approach for this conversion. We also introduce a more general framework for spectral estimation: regularized least-squares regression in reproducing kernel Hilbert spaces (RKHS). Obtained results show that the regularization framework provides an efficient approach for enhancing the generalization properties of the models.

Link functions and Matérn kernel in the estimation of reflectance spectra from RGB responses

“…We utilize a kernel-based regression model for reflectance estimation as suggested in [4][5][6]. The goal is to estimate a mapping x → q using a regression model, where we assume that x ∈ X and X is the closed and bounded subset of R k .…”

“…Our earlier research in [4][5][6] suggests that one way to increase the accuracy of estimation is via the inclusion of a priori knowledge, such as the physically feasibility of all spectral reflectance values. Here we continue this work and produce physically feasible estimations via link functions and show that the precision of the estimations is increased especially in terms of spectral shape.…”

“…Our main focus is in the comparison of the performance of link functions when combined with so-called reproducing kernel Hilbert space (RKHS) based regression models [4][5][6]. We evaluate, apparently for the first time in reflectance estimation, the Matérn kernel and copula-based link functions.…”

Estimating reflectance from multispectral camera responses based on partial least-squares regression

“…To view objects under various illumination conditions, multispectral imaging has been extensively studied to estimate spectral reflectance of object surfaces. [1][2][3][4][5][6][7] Multispectral images are usually acquired by trichromatic or monochrome cameras, accompanied by a set of color filters. The imaging process is often modeled by a linear system when the number of imaging channels is large.…”

“…5 However, the extension of polynomial responses causes overfitting and collinearity problems when the number of imaging channels is large. Heikkinen et al 6 introduced regularized polynomial modeling methods and a more general regularization framework for robust reflectance estimation.…”