Skip to main content
ABBYY Vantage uses machine learning to improve document processing. Training helps the system recognize document types and extract fields more accurately, reducing the time needed for manual verification. The system learns from corrections made during manual review.
If the system administrator has disabled training, corrections made in the Manual Review Client do not improve future processing.

​
Correcting field regions

When the extracted data on the form does not match the actual document data, verify that the field region (the area of the document image used for data extraction) is correctly defined. If necessary, go to the relevant field in the data form and select the correct region on the document image. The system then recognizes the data from the new region and displays it in the field. Once the manual review results are saved, the system uses these corrections to improve recognition quality for similar fields in future documents.

​
Correcting document types

If a document has been assigned an incorrect type, update it with the correct type. After you save the manual review results, the system becomes more accurate at assigning this document type to similar documents. For more about how Online Learning works, see Online Learning.