I need GOCR to detect characters in a 7 segment display along with dot matrix characters. I tried to GOCR as it is, to recogise the 7 segment characters, but it doesnt seem to work.. Therefore, I am thinking of adding the SSOCR engine(which deals specifically with 7 segment characters) to GOCR. Could somebody please tell me how to do this? Thank you.
No need to do that.
I've personally used GOCR to analyse fixed-font data out of old DOS captures, similar to what you want (I guess).
What is not well documented on GOCR is the options, acessible with the -m command.
2 - External DataBase (DB)
4 - Layout Analysis
8 - Don't compare unrecognized chars to recognized ones.
16 - Don't divide overlapping
32 - Don't do context detection
64 - Find Identical *exact match only)
128 - Extend DB
256 - Switch off internal engine
Just add the numbers for combinations.
I use this command line to create the database:
gocr041.exe -m 368 -p db.ocr
I use this command line to detect:
gocr041.exe -m 258 -p db.ocr
The big thing you want is to create a font file (image of each possible 7 segment chars, usualy cut and paste from the images you want to identify). Build the database with 368 (or other, try some combinations), and since it disable the internal engine, it will ask for EVERY chars, creating a solid database.
If your image is taken from video feed, I would disable the +64 of mny example :P. If your feed is near-identical to your font, you can always force up the confidence level (don't remember how, tho).
Thanks alot.Your help is really appreciated. And I would also like to know how to add an external OCR engine(like SSOCR) to GOCR. If anyone knows, please reply
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.