Hi,
In Flexicapture Studio, we have an option in pre-cognition properties: 'Thorough Extraction Mode' to force having better extraction OCR for low quality image. It's slower but can get more text in OCR in some case.
We exported the template, and use in Flexicapture Administation Station, but we don't see how to set this parameter in Flexicapture Administration Station. Some texts are recognized in Studio, but not in Administration Station.
Some one can help ?
Thanks,
Hello,
Please go to the Document Definition > Document Section properties > FlexiLayout tab > select desired option from the prerecognition mode drop-list.
Thanks, Seems OK.
Hello,
Doesn't work for me.
In Flexilayout I have an option in pre-cognition properties: 'Thorough Extraction Mode' and In FL any field are recognize.
In administration station I tried select each option from the prerecognition mode drop-list and this don't work the same like 'thorough extraction mode'.
Do you have any other solution ?
Thanks a lot
Hello,
The most popular cause for the difference in matching FC and FLS is the “Convert color and gray images to black-and-white” checkmark, which may be empty in the FLS and marked at the FC. Please make sure that they match.
Hello Ekaterina,
No difference if “Convert color and gray images to black-and-white” checkmark or not.
In FLS all field are recognize only when I have check 'Thorough Extraction Mode' but in FC I don't know what I should check or maybe write a script ?
Hello,
Could you please send us the whole image and .afl file?
Hello,
Sorry for delay in answering. We've examined the materials that you've sent and it seems to us that to understand the whole issue we also need your original project. Is it of European Invoice type? Can you describe in more details which pages have OCR differences and also attribute different results you're observing to the used settings of FLS and FC?
Hello,
Without the whole project with all its' settings we don't have a solution, only some ideas that need to be checked.
Hello,
Sorry, we cannot reproduce the described issue in all its' details. You wrote that all fields are recognized in thorough mode of FLS, but we cannot extract 7.946.527 numer in the text layer even in thorough mode.
Please sign in to leave a comment.
