E~~r~~t~on Techni for Handwri ent Preprocessing
Yan Solihin · 1996
In handwritten document image processing, a handwritten document is scuazned by a jlatbed scanner or camera arrangement. Ideally, we require to retain only the image of the handwriting without the background Unfortunuteb, in many cases, background detail is also present in the scanned image. There are several factors contributing to this. One factor is the nature of the gaper,^ it may contain color, pictures or patterns, it may be thin so handwriting on the reverse side shows through. Another factor is the handwriting implement used to produce the handwriting. Other possible factors can be defects of the scanner and improper setting of the scanning parameters. In this paper, a fuzzy based image enhancement technique to extract just the gray scale handwriting from a handwritten document image is presented The technique is based on fuzv ifthen rules applied on two fuzv sets whose parameters are estimated The technique has been tried SuccessfuDv on images of handwriting with bright-colored handwriting media, color paper, very thin paper, and paper with colorful pattern. The technique is particularly useful to the application of document processing that deal with gray level images.