PROJECT TITLE :

Color Correction Using Root-Polynomial Regression - 2015

ABSTRACT:

Cameras record three color responses (RGB) which are device dependent. Camera coordinates are mapped to a standard color space, like XYZ-useful for color measurement-by a mapping perform, e.g., the simple three×3 linear remodel (usually derived through regression). This mapping, which we have a tendency to will check with as linear color correction (LCC), has been demonstrated to figure well in the amount of studies. However, it will map RGBs to XYZs with high error. The advantage of the LCC is that it is independent of camera exposure. An alternative and probably more powerful methodology for color correction is polynomial color correction (PCC). Here, the R, G, and B values at a pixel are extended by the polynomial terms. For a given calibration training set PCC can considerably cut back the colorimetric error. However, the PCC fit depends on exposure, i.e., as exposure changes the vector of polynomial components is altered in a nonlinear way which results in hue and saturation shifts. This paper proposes a brand new polynomial-sort regression loosely connected to the concept of fractional polynomials that we call root-PCC (RPCC). Our idea is to require every term in a polynomial enlargement and take its kth root of every k-degree term. It is simple to indicate terms outlined during this manner scale with exposure. RPCC is a easy (low complexity) extension of LCC. The experiments presented during this paper demonstrate that RPCC enhances color correction performance on real and synthetic data.