Dense Sampling of 3D Color Transfer Functions Using HDR Photography

Marcel Heinz, Guido Brunnett; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2015, pp. 25-32

To apply brightness and color adjustments to projected images,the color transfer function (CTF) of the projector has to be known. We propose a novel approach to determine the CTF using a high sampling density, which is suitable for modern DLP projectors working with color wheels with additional primaries. Our approach is based on the principle of measuring patterns consisting thousands of color samples at once, using a DSLR camera and high dynamic range photography. To ensure high accuracy, additional correction patterns are introduced to compensate for the influence of the dynamic background light caused by displaying the patterns itself. Furthermore, several permutations of the samples in the patterns are captured to address spatial variances of both the projector and the camera. We show that our method achieves comparable accuracy to existing methods, but is one to two orders of magnitude faster. A 64^3 sampling of the CTF can be acquired in a few hours, compared to several weeks that sequential spot measurements would take. Additionally, we demonstrate that a different configuration of our method can be used to capture 17^3 samples extremely fast, indicating the applicability for cases where sparse sampling is sufficient.