Document Extraction Character Confidence

  • Last Post 03 August 2018
semicolon87 posted this 31 July 2018

Good day,

I am busy with document extraction, and I would like to also extract the Confidence score per field.

Currently there is a Paramater in the ICharacterParams interface with the name Confidence:


Confidence level for a batch type. Returns a value from 0 to 100.


Important! Using this property requires storing some additional data, which may increase the size of a document and additional disk space. By default, no additional data are saved and the property is disabled. For the correct operation of the property, set the registry key [HKEY_CURRENT_USER\Software\ABBYY\FlexiCapture\12.0\Controller] "ForceSaveConfidenceLevel" = "true". Instead of HKEY_CURRENT_USER, select the branch of the user account that will recognize documents.

ICharacterParams Link

This only gives you the Batch Confidence, is there any way to get the field's individual confidence?


Helen Osetrova posted this 03 August 2018



If your question is related to ABBYY FlexiCapture desktop product, please ask it on This forum works for ABBYY SDK products only.


If you are working with ABBYY FlexiCapture Engine, there are a few options to get the confidence level for the words and characters:

• for a character, please find it in the CharConfidence property of the CharParams and RecognizedCharacterInfo objects,

• for a word, please check the WordConfidence property of the RecognizedWordInfo object.