Stereo Retinex

The Retinex algorithm for lightness and color constancy is extended here to include 3-dimensional spatial information reconstructed from a stereo image. Tests show this modification improves retinex. A key aspect of traditional Retinex is that, within each color channel, it makes local spatial comparisons of intensity. In particular, intensities ratios are computed between neighboring spatial locations. Retinex assumes that a large ratio indicates a change in surface reflectance, not a change in incident illumination; however, this assumption is often violated in 3-dimensional scenes, where an abrupt change in surface orientation can lead to a significant change in illumination. Stereo Retinex uses a stereoderived depth map to locate abrupt changes in surface orientation so that it can avoid making spatial comparisons at locations where an illumination change would otherwise be mistaken for a reflectance change. The stereo retinex algorithm builds upon the multi-resolution implementation of retinex known as McCann99. It is modified to use the 3-dimensional edge information derived from stereo images. The edge map is propagated through all levels of the multi-resolution pyramid so that at each level spatial comparisons are only made between locations lying on approximately the same plane. Experiments on real images show that Stereo Retinex performs significantly better than unmodified McCann99 retinex when evaluated in terms of accuracy with which correct surface object colors are estimated.