Simple table definition in templates

  • Last Post 14 March 2014
microsign posted this 13 March 2014


I am fairly new to the ABBYY OCR SDK. I am using the demo-project "Auto-Learning Technology Demo" as a stepping stone for a custom template designer. I'd like to know how I can program table/column/(repeating) row recognition/auto-detection, because I cannot find this in the documentation (or I'm not interpreting it correct).

Can anyone give me some pointers or a code snippet as to how I can implement this? The current functionality is as follows:

  1. User starts a new definition
  2. User chooses images
  3. User draws boxes where the data and references are

Step 3 should be expanded with the possibility to select and auto-detect tables, rows and/or columns.

Thanks in advance!

Order By: Standard | Newest | Votes
SDK_support posted this 14 March 2014

Do we understand correctly that you are using FlexiCapture Engine? It is possible to train the document definition to find only simple fields during the document creation via training.

microsign posted this 14 March 2014

That is correct. Do you mean it is only possible to find simple fields with the SDK, and that complex fields cannot be done thorugh the SDK and should be defined with the FC Template Editor?

SDK_support posted this 14 March 2014

I meant that it is only possible to create template via auto training that will find only text fields. In general it is of course possible to process complex fields via FlexiCapture Engine, but for that purposes you need to create appropriate template in FC Template Editor.