Nonuniform Lattice Regression for Modeling the Camera Imaging Pipeline

Hai Ting Lin, Zheng Lu, Seon Joo Kim, Michael S. Brown

Abstract: We describe a method to construct a sparse lookup table (LUT) that is effective in modeling the camera imaging pipeline that maps a RAW camera values to their sRGB output. This work builds on the recent in-camera color processing model proposed by Kim et al. [1] that included a 3D gamut-mapping function. The major drawback in [1] is the high computational cost of the 3D mapping function that uses radial basis functions (RBF) involving several thousand control points. We show how to construct a LUT using a novel nonuniform lattice regression method that adapts the LUT lattice to better fit the 3D gamut-mapping function. Our method offers not only a performance speedup of an order of magnitude faster than RBF, but also a compact mechanism to describe the imaging pipeline.