An effective grayscale conversion with applications to image enhancement

Abstract

Recent efforts to accurately integrate properties of human visual perception have fueled an impressive improvement not only in digital photography but also in color manipulation tools. Color is one of the main attributes that influences human visual system, being ultimately the result of our experience and perception. However, several important applications such as substitution systems for visually impaired people, computerized tomography, printed textbooks and aesthetical stylization still require a reliable grayscale representation of images. Mapping three dimensional color information onto a single dimension while still preserving the original appearance, contrast and finest details is not a trivial task. Standard monochromatic transformations, found in commercial image editing software, neglect the color distribution, being unable to conserve in many cases the discriminability of the original chromatic contrast. Consequently, isoluminant colors are mapped to similar grayscale levels increasing the ambiguity in the decolorized image version. Although luminance is a definite value, our subjective interpretation that measures apparent brightness is very important in image decolorization process.