Dynamic range expansion of single images using intensity-invariant patch correspondences.

We propose a new method(1) for expanding the dynamic range (DR) of a single standard dynamic range (SDR) image by exploiting the self-similarity property of natural images and scene lighting variations. Our method first uses a proposed robust distance metric to match ill-exposed (i.e., over-or under-exposed) patches of the input image with well-exposed patches from the same image; it then uses details from these matched patches to restore the ill-exposed patches. We model this restoration problem as global optimization where the energy function penalizes the difference between the matching patches. Since this system is underdetermined we use natural image statistics for well-posedness. Our results demonstrate improvement over the state-of-the-art. Our results show improvement compared to the state-of-the-art.