When using the LF Connector test utility to capture OCR zones, we saw that it would misread the number zero as the letter O, or the letter O would be misread as zero.
Is there a way to improve the "accuracy" of the OCR task?
When using the LF Connector test utility to capture OCR zones, we saw that it would misread the number zero as the letter O, or the letter O would be misread as zero.
Is there a way to improve the "accuracy" of the OCR task?
Hi Kyle,
There are a number of OCR settings you can tweak in the Connector profile. Check out https://www.laserfiche.com/support/webhelp/Laserfiche/10/en-US/administration/#../Subsystems/Connector/Content/Editing-XML-Profiles.htm#OCRImageEnhancementandLanguageSettings for the full list.
You can even tweak this in the test utility if you just want to quickly try various combinations out by updating the OCRParameters.xml file in the test utility directory.
Justin,
Can you give a write up of what each OCR parameter can be adjusted to do, or what range is available? And if there is a range for a parameter, what does it do when you increase of decrease the value? Something like this?
Note: after re-reading the help file, I found most of the information I was looking for. And I then wrote it up in a file like seen below.