Point Cloud Color Constancy

In this paper, we present Point Cloud Color Constancy, in short PCCC, an illumination chromaticity estimation algorithm exploiting a point cloud. We leverage the depth information captured by the time-of-flight (ToF) sensor mounted rigidly with the RGB sensor, and form a 6D cloud where each point contains the coordinates and RGB intensities, noted as (x,y,z,r,g,b). PCCC applies the PointNet architecture to the color constancy problem, deriving the illumination vector point-wise and then making a global decision about the global illumination chromaticity. On two popular RGB-D datasets, which we extend with illumination information, as well as on a novel benchmark, PCCC obtains lower error than the state-of-the-art algorithms. Our method is simple and fast, requiring merely 16 x 16-size input and reaching speed over 140 fps (CPU time), including the cost of building the point cloud and net inference.

Automated summary

In this paper, we propose PCCC, an algorithm for illumination chromaticity estimation using point clouds. By leveraging depth information from a ToF sensor and RGB intensities, PCCC applies the PointNet architecture to derive the illumination vector and make a global decision about the global illumination chromaticity. PCCC outperforms state-of-the-art algorithms on popular RGB-D datasets and a novel benchmark, with a simple and fast method that requires a small input size.