Root-Polynomial Color Homography Color Correction

Homographies are at the heart of computer vision and they are used in geometric camera calibration, image registration, and stereo vision and other tasks. In geometric computer vision, two images of the same 3D plane captured in two different viewing locations are related by a planar (2D) homography. Recent work showed that the concept of a planar homography mapping can be applied to shading-invariant color correction. In this paper, we extend the color homography color correction idea by incorporating higher order root-polynomial terms into the color correction problem formulation. Our experiments show that our new shading-invariant color correction method can obtain yet more accurate and stable performance compared with the previous 2D color homography method.