Community

Improve Pre-recognization?

Avatar
Written by Permanently deleted user
  • 3
Hello all,

I am new to flexicapture.
I am confused by pre-recognization. The static texts and regular expression match keep on failing as pre-recognization fails.

Is there a way to change/improve image settings like contrast when adding them without going through scanning station? (I only found the auto-rotation,despeckle ones)
Also, is there a way to improve or limit the character sets for pre-reorganization? (Like this image won't have character '*', Not select language nor mode)

Thanks,
Yun

Was this article helpful?

0 out of 0 found this helpful

Comments

3 comments

Sort by Date Votes
  • Avatar
    Permanently deleted user
    Hello,

    Pre-recognition options can be specified during image import on
    -Scaning station
    -When you manually add images on verification station or project setup station
    -If you are using hot folders, then pre-recognition stetting can be specified in image profiles

    Prerecgnition settings are the same everuwhere:



    If some elements like static texts are not found, then you should increase error procentage in element properties in Flexilayout studio or use an alternative approach to capture text.

    Regards
    Timour
























    0
  • Avatar
    Permanently deleted user
    Hi Timour,

    Thank you for your response.

    I have tried those and the situation I have is the pics are taken by cell phones rather than a scanner which has lower image quality.
    In this case, even with the Thorough mode for pre-recognition, the text recognition is still not performing very well.
    The problem with error procentage increase is the static texts I am looking for are very short and it would very likely to find other words instead.

    Is it possible to limit the pre-recognition characters? Like never recognize a character as '?'.
    Also, for the offset value set inside relationships, will it automatically proportioned when recognizing a bigger image?

    Thanks,
    Yun Zhou
    0
  • Avatar
    Permanently deleted user
    Hello,

    Is it possible to limit the pre-recognition characters? Like never recognize a character as '?'.
    Such functionality is not implemented.

    Also, for the offset value set inside relationships, will it automatically proportioned when recognizing a bigger image?
    No, if offset is specified in dots or pixels then it will not be proportioned according to image size.
    In order to handle such cases you need to use advanced pre-search relations instead of relations dialogue, and use found elements to specify offset values.

    Code in pre-search relations should look similar to the follwoing:

    //relations and offsets
    leftof:ELEMENT1.Rect.right+ELEMENT1.Rect.Height*2;
    rightof:ELEMENT1.Rect.left+ELEMENT1.Rect.Height*2;
    above:ELEMENT1.Rect.bottom+ELEMENT1.Rect.Height*2;
    below:ELEMENT1.Rect.Top+ELEMENT1.Rect.Height*2;
    nearest:rightof:ELEMENT1;

    As you can see, in this case search relations are specified relatively to ELEMENT1. Image size in this case won't matter.

    You can find more information about how to use advanced pre-search relations and flexilayout language in FlexiLayout help file.


    Best Regards,
    Tiomour








    0

Please sign in to leave a comment.

Didn't find what you were looking for?

New post