The quantitative characterization of structures or textures in images relates scales in the spatial domain with those in color space. The image enhancement shown here is result of an information-based method, in which the scale of the color space is optimized according to the structural content of the image. The "information-bearing color values" are determined and used for adaption of the color scale (usually an inhomogeneous, nonlinear partition).

The image enhancement runs fully automatic, i.e. no supervision is needed. However, it is possible to provide one parameter, which defines the spatial scale of expected structures. Note, the color rescaling does not lead in any case to aesthetically improved solutions (even though this is often the case!). It proves rather as a useful preprocessing step for algorithmic processing of images in segmentation or classifcation tasks. It is especially helpful for discriminating structures or a quantitative assessment, e.g. in medical imaging.

Olympic Stadium Munich:

Allianz Arena Munich:

Statue of Martin Luther, Dresden: