コミュニティ

Wrong recognition is happening with higher confidence level

Hi,

I am a user of ABBYY Flexicapture 11 Distributed. Many times it is happening that ABBYY is reading some characters on the document wrongly but with high confidence level, so because of it those characters are coming in black and getting skipped from character verification. If I decrease the threshold percentage for red highlighting then more characters are coming for verification which is obviously not feasible.

Anyone have worked on such scenario? can anyone give a solution for this problem?

Thanks in advance,
Rahul

この記事は役に立ちましたか?

1人中0人がこの記事が役に立ったと言っています

コメント

1件のコメント

  • Avatar
    Permanently deleted user

    Hello,

    Due to some factors such as image quality or fonts used in the original document, FlexiCapture recognition could have systematic errors. If the recognition accuracy setting in the verification tab of the field properties gives too many symbols to verify, you may try to increase the recognition accuracy and decrease the error level by using special data types in the Data tabs in the field properties or apply the auto-correction script.

    To give you a better advice, we need to look at the source images and recognition results.

    0

サインインしてコメントを残してください。