Hi Vicky, Well, yes. I am actually trying to first localize text in images meaning finding text blocks within the image. I already put some algorithms in place which do the job for me. But i want to see if tesseract OCR also doing it before actually recognizing the characters. If it is the case, i want to compare my results with tesseract one.
Also as text localization will vary from application to application and we may need to set different parameters in different scenarios. So i just thought it would be better if i optimize text localization for my application before giving those text blocks to tesseract OCR.. I hope it helps. Thanks and Regards Parmeet On Wed, May 25, 2011 at 10:26 PM, Vicky Budhiraja < [email protected]> wrote: > Hi Parmeet, > > Sorry, I was not able to understand your question(s). Are you looking for a > method that can find text within graphics? Like text in Video frames or > Stylized text in a print-ad etc.? > > Best Regards, > Vicky > > -- > Vicky Budhiraja > http://www.sitarasoft.com/ > > -----Original Message----- > From: [email protected] [mailto: > [email protected]] > On Behalf Of Parmeet > Sent: Wednesday, May 25, 2011 16:51 > To: tesseract-ocr > Subject: Localized text in images > > Hi all, > > I am working on recognizing text from device images. As i go through > papers, i find that it is always advisable to find text regions(text > localization) and then give it to any standard OCR engine to increase > the recognition accuracy. > > I wonder, if Tesseract OCR itself try to localize the text first > before OCRing. If it is the case, can we somehow get the intermediate > result of localized text from tesseract OCR, so that i can compare it > with text localization results i obtained using other algorithms. This > way i want to make sure that the results of OCR i am getting are the > optimal one. > > Also same is i wanted to ask for other intermediate results like > binarization, skew correction etc. Would there be a way to get all the > intermediate results before tesseract OCR actually work on recognizing > characters? > > Thanks and Regards > Parmeet > > -- > 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 > -- Parmeet https://sites.google.com/site/bhatiaparmeet/ -- 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
