Nonlinear Unsharp Masking for the enhancement of document images

A novel operator for the enhancement of the quality of document images is presented in this paper. This operator, which is a quadratic one, is based on the Unsharp Masking (UM) technique, but it is able to limit noise amplification because every pixel of the processed image depends upon a large portion of the input image; in the same time a good response on details is obtained. A formal description of the operator's response to noise is also presented.