If you have any way to detect or specify regions of same background/foreground colors, Otsu will work fine for each such region. If you don't, then for the entire image apply any *adaptive* binarization (even one of the simplest ones) and it should work fine. First you can try Niblack. All this holds true if your other images do not differ much from what you've shown here in the forum.
Warm regards, Dmitri Silaev www.CustomOCR.com On Thu, Nov 29, 2012 at 1:07 PM, sascha4j <[email protected]> wrote: > thank you for your answer i will take a look at your example and the > leptonica library > > yesterday i looked for binarization algorithms. of course otsu is in this > example no option. an adaptive algorithm is needed. i tried several like > sauvola or niblack. > > at the moment best seems for me NICK > > described in this document > > > http://www.math-info.univ-paris5.fr/~vincent/articles/DRR_nick_binarization_09.pdf > > > > > > Am Mittwoch, 28. November 2012 11:10:07 UTC+1 schrieb sascha4j: > >> i have trouble to ocr an image like in the attachment. >> >> only the word text is recognized. >> >> i tried several binarization algorithms, but without success. >> >> does it make sense to binarize the image ? or has tesseract it's own >> binarization? >> >> any hints would be nice. >> >> greetings >> sascha4j >> >> >> >> > -- > You received this message because you are subscribed to the Google > Groups "tesseract-ocr" group. > To post to this group, send email to [email protected] > To unsubscribe from this group, send email to > [email protected] > For more options, visit this group at > http://groups.google.com/group/tesseract-ocr?hl=en > -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/tesseract-ocr?hl=en
