How to classify images using FlexiCapture 12 ?


For a customer we wanted to process invoices coming from mailboxes, we managed to set up an image import profile using the Microsoft Graph Mail API. Our problem is that in addition to invoices, we sometimes receive other images, such as vehicle images or vehicle technical control images (roadworthiness), and from these images we don’t need to extract data, we just need to classify them to know if they are vehicle images or vehicle technical control images. How to set up this using FlexiCapture 12 ? Should we create a classifier or is there another solution ?

Was this article helpful?

1 out of 1 found this helpful


1 comment

  • Avatar
    Scott Chau


    A better option is to use some sort of email filter to move the email with invoice to the correct location then have FlexiCapture grab from that email folder.  The reason for this, if you let FlexiCapture do the "sorting" all those pages will take a page count too, which could exhaust your page count very quickly.

    Assuming you have an unlimited page count or purchase a license page count that accounted for non-invoices, the invoice project should do a good job and figuring out what is an Invoice or not.  Basically there are some logic where it looks for the Invoice No, Total amount, etc and if it doesn't see them, won't consider them an invoice.  Now lets say you have a situation where you have a batch and its a mix of all these document, then you would need to setup a classification project with a custom workflow stage.  Basically you would have to script it to assemble the document based on some logic like if the invoice number was the same.


Please sign in to leave a comment.