There is a possible bug in the TrainUserPatterns or at least in the code sample "User Pattern Training Utility"
In the sample the Pattern Training always skips the last symbol in a word. This is a problem because closing brackets often only appear at the end of a word and therefore are never trained.
I don't know if the TrainUserPatterns works the same as the demo because it only asks to enter a character for unkown character images whereas the sample asks for every letter (frame).
コメント
1件のコメント
Hello,
Please sorry for possible delay with response! I have just processed one English document using FineReader Engine 11 (build 11.1.19.82) Demo Tools → User Pattern Training Utility and it seems everything works well enough. During training some characters can be skipped if the program considers that the character is recognized confidently, especially this often happens by the end of the document, when most of the alphabet characters are already trained. To clarify the issue please provide the following information to SDK_Support@abbyy.com:
1. The build number of your FineReader Engine distribution kit.
2. The document, which you use for training.
3. What language do you use for training?
サインインしてコメントを残してください。