Targeted attribute transformation involves obtaining a digital image of the skin, selecting an individual skin attribute (an imperfection, such as a blemishes, pores, or wrinkles), extracting the individual skin attribute from the digital image by linear digital filtering on all the color channels, to obtain a contrast map in all three channels of the individual attribute, and adding or subtracting a fraction of the contrast map to the initial image, to obtain the skin image with transformed selected skin attribute.