Have you ever been working on a zonal OCR project only to discover that the scanned documents were too different from the original document for OCR to be of any practical use? Some differences such as resolution and skew angle can be algorithmically corrected, but what if the zones themselves change in size? What if the data you want to capture is tabular in nature and has an unknown number of rows? Unstructured forms and scanned documents such as invoices, bills and tabs are notoriously difficult, if not impossible, to process on a large scale within an automated system. LEADTOOLS introduced its Invoice Recognition and Processing SDK technology in Version 19 to address this very problem.