Contrast-Gain-Based Visual Tone Mapping for Digital Photo Prints

Abstract

With the popularization of digital photo printing, there is a demand for print image quality that approaches the quality of silver halide photographs. One way to achieve this is through automatic correction of image luminance/tone. The conventional approach based on spatially invariant mapping often causes undesired luminance changes when enhancing contrast; moreover, luminance changes are also accompanied by changes in contrast. This paper proposes a novel algorithm for the enhancement of the contrast and lightness using spatially variant mapping to achieve high-quality printing of images photographed using digital cameras. First, we define a "contrast-gain" function to quantitatively evaluate the visual contrast that results from tone mapping. Based on this function, the proposed tone mapping algorithm enables independent control of the lightness and contrast through spatially variant processing. Next, we formalize a "shadow-up tone curve," which produces an effect similar to auxiliary lighting during photography. Our contrast-gain-based visual tone mapping method, which uses the shadow-up tone curve in the algorithm, has none of the unnaturalness that arises with a conventional single tone curve. Furthermore, it provides an extremely natural effect similar to that in images photographed by a professional photographer using auxiliary light provided by a reflection board.