Abstract: An document image that is the source of Optical Character Recognition (OCR) output is displayed so that a user can select a region of the displayed document image. When the region is selected, text of the OCR output corresponding to the selected region is submitted as an input to a search engine.