Entire adjustment of tonal values using hierarchical interpolation

We present a new approach for adjusting the entire tonal values and other appearance parameters in an image. In our framework, the user just indicates regions of interest by drawing some rough strokes and then adjusts the hue, brightness, color, and other parameters in these regions. A hierarchical interpolation formulation is used to propagate the initial adjustment to the entire image by enforcing the policy that similar edits are applied to spatially-close regions of similar appearance. We show that our hierarchical clustering is a novel and efficient method for approximately solving the propagation process. The propagation approach developed in this paper is a general one, and also used to support specifying a wide range of complex images. Furthermore, we show that for high-resolution images the proposed method outperforms prior methods which just take a few seconds.