A Multiresolution Approach to Recovering Colors and Details of Clipped Image Regions

In regard with extrinsic and intrinsic parameters of a camera system, either factory or man-made settings may degrade the quality of the output image. Specifically, inappropriate exposure control may unintentionally cause the resultant image containing over-exposure regions and/or saturated color channel(s). Such an effect is the so-called clipped pixels or clipped image regions. In this paper, we propose a multiresolution approach to recovering colors and details of clipped image regions. In the proposed scheme, the first step is to decompose an input image into a smooth image and the detail images at different scales. As for the smooth image, we propose a new desaturation method, which attempts to restore and recover the intensities and colors of the clipped image regions. As for the detail images, we propose the use of the detail enhancement method, which makes the extracted image details preserved or enhanced as the case may be. Moreover, a series of experiments have shown that the proposed scheme is indeed effective in recovering the colors and details of the clipped image regions.