Semi-structured documents

In many cases you know the context of the information on a document, however its location is shifted. You may want to know the total amount due or company name, but the document does not have a fixed layout and therefore these data fields can be anywhere on the page. Invoices are a good example of such document. LeapReader is designed to address such documents.

The system goes through a five stage process to provide you with the best recognition possible. Starting with image processing, our engine removes noise, deskew's the document and enhances the image. Later the system normalizes, segments and extracts the information and finally it validates and ranks the data fields by confidence level ensuring that both speed and accuracy are addressed.