Identification and removal of advertisements from yellow page documents

OCRing fails to deliver the information embaded in a yellow page document. Such failure stems from the fact that a yellow page document includes creative advertisements, multiple columns, and decorative graphics. In this research effort we introduce a set of algorithms that enables us to identify and remove advertisements from a scanned yellow page document, Removal of advertisements is a major step in paving the way for successful OCRing of the yellow pages. The scanned image is a gray scaled image with 256 gray levels and the resolution of 3300 x 4400. The experimental test shows 98% correct identification and removal of advertisements from the image.