Hi all, I have several scanned documents that have a lot of noise between lines and between words. Tesseract fails to ignore them and it either includes them in the next character or makes them a separate character, often a dot or comma. I attached an image that shows some of that noise.
I am using the latest SVN version of Tesseract 3.03. Tesseract 3.02 does slightly better at ignoring the noise. Now my questions are: 1. What are the configuration parameters (maybe also hard coded constants) inside Tesseract that affect the noise vs. good blob classification? 2. Is there a way to define a minimum number of pixels or dimensions for a connected component? 3. Is there a way to limit the scaling of a blob, so that it won't get matched to a character prototype? I already found the configuration parameters heavy_noise_reduction and textord_noise_hfract, but heavy_noise_reduction gives me bad results and by teaking textord_noise_reduction I can get better results, but they still aren't satisfying. Maybe there's a better alternative in the code that I can't find. Regards, Paul -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/129bbd4f-ff94-40a0-85a8-a8a9740bbf1e%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
