When working with FlexiCapture SDK there are different ways to implement the Training functionality.
If your plan is to initiate FlexiCapture Training during the actual document workflow, this could be achieved by implementing a verification step.
The verification interface, which has to be created on the customer´s side, will need be able to capture the the X- and Y- coordinates on an Image. This way you can provide FlexiCapture the exact position of the missing field information, so that FlexiCapture can trigger its advanced training functionality. Just by providing the fields position, FlexiCapture can start to understand the nature of the field like its data type, surroundings, patterns, etc.
Depending on your solution´s infrastructure, there are different ways to implement the verification step.
For example, if building a web based application, you could use jQuery to capture the relative coordinates on the image. Link
Once the required coordinates are stored, you can use them to create a new region object that FC SDK can use.
It is also possible to initiate FC Training without using a verification interface but with the fulltext OCR results.
In the Invoices code sample of FC SDK, a helper function is used to determine the exact coordinates for fieldtraining:
Basically by using “FieldsHelper.GetRectFromPageByText” we search the information in the OCR fulltext, and let FC SDK create the region around it to find the exact X and Y coordinates.
Please sign in to leave a comment.