摘要 |
<p>Vouchers (such as credit card vouchers) normally have a fixed layout, with areas on which fixed information is printed and a plurality of zones in which information has been entered manually; a zone normally consists of several sub-zones, one per character. To read such vouchers, a scanner 11 and image generating means 12 scan the voucher and create a voucher image file. A forms library 13 stores a plurality of form type images, including images in both a correct and an inverted orientation for each form type. A form recognition unit 15 compares the voucher image with the form type images to identify the form type of the voucher. Form removal means 16 remove the fixed form information from the voucher image. A character recognition unit 25 identifies the character (if any) in each sub-zone. Validation means 27 and 28 then validate the information so read, using a monitor 29 and a keyboard 14 for operator control. <IMAGE></p> |