Confidence Level

  • Last Post 17 April 2019
abbyyuser posted this 10 March 2017


On my project station, it was stated that the Confidence Level is 96% (133 of 138).
This document has 4 pages, with details as follows:
94% (78 of 83)
0% (0 of 0)
0% (0 of 0)
0% (0 of 0)

When I extract the following value from the Document, they are as follows:
RecognizedSymbolsCount = 83
SymbolsForVerificationCount = 5
TotalSymbolsCount = 83
UncertainSymbolsCount = 5

How does "96% (133 of 138)" figure come about?



Order By: Standard | Newest | Votes
Ekaterina posted this 13 March 2017


In the TotalSymbolsCount field, the program shows the count of all symbols, punctuation marks, spaces, checkmarks, groups of checkmarks and fields. The UncertainSymbols field is the count of uncertainly recognized symbols. So as the statistic is based not only on symbols, but also on fields, it will show different confidence value.

abbyyuser posted this 16 March 2017

Hi Ekaterina,

Is there any method in the Document object to retrieve the 96% (133 of 138) values?
The current solution we have built for our client has a feature to send email to notify them on the confidence level upon document recognition.

Obviously, we can't tell them that the confidence level is 94% (78 of 83), when the project station shows a different value.



Ekaterina posted this 16 March 2017


There is not such method. posted this 15 March 2019

Hi Ekaterina,


We are building a Proof of Concept for our client and we are planning to use abbyy for reading and extracting data from PDF files.

We have selected ABBYY Flexi Capture version 12 as our choice for OCR but are doing this PoC using a trial version for our initial demo.

One of the requirements of our client is to extract a confidence level at a field level and at a document level. Is it possible to fulfill this requirement in the trial version?


Ekaterina posted this 17 April 2019


Yes, it is possible.