Nonlinear digital filters for beautifying facial images in multimedia systems

Nonlinear digital filters for new application in multimedia systems, that is beautification of face images, are proposed. These filters can remove undesirable skin components such as wrinkles and spots on face images, and accordingly make the skin look smooth and clear without affecting other principal components in the images. Three types of nonlinear filters are proposed: one is quite a simple one which is effective for removing roughness and small spots on skin and can be easily realized by simple hardware, one is an extended one which is a little more complicated but can smooth the roughness and small spots more naturally, and the other more complicated one which can preserve the natural roughness on skin and can remove relatively large spots on skin. Computer simulations verify the high performance of these filters.