or do you advise to use an other librery like opengl to get char images? Thanks for time.
2011/10/5 merve t <[email protected]> > Hello, > Thanks for your encouraging response, i was about to give up to try. > > I am using PSM_AUTO > > What is the most elegant way of getting images char by char? > > I tried to access tesseract's interior structures a few days ago, but get > "this is protected" error. > > I know there is a GetImage method in PageIterator but i do not know if it > works that i generate a page iterator using entire image. > > > > > 2011/10/5 Saurabh Gandhi <[email protected]> > >> Sorry, the reply I gave earlier was based on a misinterpretation of your >> problem....why dont you try to segment the characters yourself and then hand >> over every character to Tesseract for recognition in PSM_CHAR mode..that >> should solve the matter...which mode are you using Tesseract in currently? >> >> -- >> Regards, >> Saurabh Gandhi >> >> >> >> >> >> On Tue, Oct 4, 2011 at 8:03 PM, merve t <[email protected]> wrote: >> >>> Gandhi, thanks for answer, i do not have a unicharambigs file in >>> training, i did not generate one while training. >>> >>> Parmar, thanks for answer, there is no black or gray pixels between l and >>> t, i attached the file. >>> >>> 2011/10/4 Sandeep Parmar <[email protected]> >>> >>>> Zoom ur image and check whether the character 'I' and 't' are joined or >>>> not. I guess pixels of both characters are joined thats why Tesseract >>>> spells >>>> it 'H'. >>>> >>>> Please confirm >>>> >>>> Regards >>>> Sandeep >>>> >>>> On Tue, Oct 4, 2011 at 2:53 PM, merve t <[email protected]> wrote: >>>> >>>>> Hello, >>>>> I think my problem is related to this. >>>>> But a little different >>>>> >>>>> for example in my image there is >>>>> >>>>> "lt" >>>>> >>>>> and tesseract recognizes this as "H". >>>>> >>>>> i do not want tesseract to combine blobs, i want it recognize blobs >>>>> individually. >>>>> >>>>> What should i do? >>>>> >>>>> >>>>> 2011/10/4 Saurabh Gandhi <[email protected]> >>>>> >>>>>> Hi Rakesh, >>>>>> >>>>>> We did it in chopper.cpp within chop_word_main()... >>>>>> >>>>>> -- >>>>>> Regards, >>>>>> Saurabh Gandhi >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> On Tue, Oct 4, 2011 at 10:49 AM, Rakesh A <[email protected]>wrote: >>>>>> >>>>>>> Hi Saurabh, >>>>>>> >>>>>>> should it be set in wordrec.cpp or wordrec.h? or do you suggest >>>>>>> tessbox.cpp? >>>>>>> Which is the cleanest way of doing it? Can it be set in a config >>>>>>> file? >>>>>>> >>>>>>> Thanks a lot >>>>>>> >>>>>>> >>>>>>> On Mon, Sep 26, 2011 at 8:48 PM, Saurabh Gandhi < >>>>>>> [email protected]> wrote: >>>>>>> >>>>>>>> By setting chop_enable as zero, we were able to disable the >>>>>>>> splitting of a single character. >>>>>>>> Not sure if that's the correct approach, may be Dmitri can throw >>>>>>>> more light...:) >>>>>>>> >>>>>>>> -- >>>>>>>> Regards, >>>>>>>> Saurabh Gandhi >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Tue, Sep 27, 2011 at 12:21 AM, speeder <[email protected]>wrote: >>>>>>>> >>>>>>>>> I did not knew this feature existed... and ask the same, it is >>>>>>>>> causing some annoying bugs on my software (namely, I am using ONLY >>>>>>>>> digits, + >>>>>>>>> and - characters, yet sometimes a - become -- or even --- and >>>>>>>>> sometimes a >>>>>>>>> number become two, with 8 sometimes becoming 11 for example, or 14 or >>>>>>>>> something like that). >>>>>>>>> >>>>>>>>> Also, avoiding joins would be nice too (avoid 14 to become 8 for >>>>>>>>> example). >>>>>>>>> >>>>>>>>> >>>>>>>>> On Mon, Sep 26, 2011 at 10:55 AM, iqyush <[email protected]> wrote: >>>>>>>>> >>>>>>>>>> Hi all: >>>>>>>>>> The attached BMP should be recognized by tesseract OCR 3.0.1 >>>>>>>>>> as >>>>>>>>>> TULOMSAS 11-2-06377 MG 11 09 ER9 T 018 >>>>>>>>>> but it's recognized as >>>>>>>>>> TU*2*LOMSAS 11*I*-2-06377 MG 11 09 ER9 T 018 >>>>>>>>>> I think the char "L" is splitted into and recognized as 2L; >>>>>>>>>> and the second char "1" is splitted into and reognized as "1I". >>>>>>>>>> >>>>>>>>>> In my application the font size at a line is same, so OCR script >>>>>>>>>> feature is not needed, >>>>>>>>>> so can any one tell me how to disable the multi split feature from >>>>>>>>>> configure file? >>>>>>>>>> >>>>>>>>>> B.R. Aiqing Yu >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> -- >>>>>>>>>> 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 >>>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> 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 >>>>>>> >>>>>> >>>>>> -- >>>>>> 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 >>>>> >>>> >>>> -- >>>> 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 >>> >> >> -- >> 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
