Displaying a graphics such as for example a diagram or a map on a display unit of a portable device requires a reduction in the number of pixels. In each case this requires a decomposition of the graphic into a number of blocks. To do this each block is usually subjected to what is known as "subsampling. In such cases the semantically important contours disappear at least partly so that the content of the graphic is lost. To resolve this problem to a method is proposed which for a specific application assigns a weighting to the colors underlying the contours. For the reduction in dominant color is determined depending on the result of the weighting and this is simulated in a further procedural step from a set of specified basic and pseudo colors. In this way the semantically important contours of a graphic and thereby the readability are retained despite the fact that reduction has been undertaken.