:) I hear you, I do. From the input image perspective I looked at the 2 Kaliko words and they match at the pixel level. Some part of Tesseract's internal classification system is falling over a bit here, I cannot tell you why, I can only say that OCR is not a perfect technology. It works on machine learning and classification of probabilities. What seems simple to us is not necessarily so to the OCR engine. It could be that by the time it gets to the 2nd Kaliko it thinks it has improved the 'l' recognition from the previous text to be a '1' - Tesseract has this adaptive classifier - but I am only guessing, maybe others have better ideas.
On 8 January 2015 at 15:24, Gokcer Gunes <[email protected]> wrote: > and also why is that? i mean at 3. line there is both KalikoImage which > is same,but first is correctly translated to text and second one is with > one mistake > > 2015-01-08 17:17 GMT+02:00 Allistair <[email protected]>: > >> It works for me with just the threshold and no scaling (however, there is >> an error in the 2nd Kaliko where the l is misrecognised as a 1): >> >> https://dl.dropboxusercontent.com/u/523401/codethres.png >> >> --- >> >> Memorystream memstream = new MemoryStream(); >> >> bitmap.Save(memstream, System.Drawing.Imaging.ImageFormat.Png); >> >> Kalikolmage imgklk = new Ka1ikoImage(memstream); >> >> >> On 8 January 2015 at 14:22, Gokcer Gunes <[email protected]> wrote: >> >>> Yeah with more size it works better and better but it takes more time >>> also thats why im only scaling 2x around.And can i ask did you try with >>> only threshold or threshold+scaling >>> >>> 2015-01-08 16:16 GMT+02:00 Allistair <[email protected]>: >>> >>>> Oh, I also scaled the image up to 2000px wide. Larger images often work >>>> better with Tesseract so a pre-filter or sampling up can help. >>>> >>>> On 8 January 2015 at 14:16, Allistair <[email protected]> wrote: >>>> >>>>> This is the image I am providing to Tesseract >>>>> >>>>> https://dl.dropboxusercontent.com/u/523401/code.png >>>>> >>>>> It's your image with a quick threshold performed in Photoshop. Notice >>>>> that you need to ensure the threshold makes all text black. >>>>> >>>>> On 8 January 2015 at 13:47, Gokcer Gunes <[email protected]> wrote: >>>>> >>>>>> i tried with threshold filters too but i didnt get good results,best >>>>>> result for me was this way.Can i ask what threshould number you used for >>>>>> this images. >>>>>> >>>>>> 2015-01-08 14:30 GMT+02:00 Allistair <[email protected]>: >>>>>> >>>>>>> Well, I must say I am unsure why one image works for that line and >>>>>>> not the other. Clearly the addition of other lines of text is causing >>>>>>> Tesseract's classification system to go slightly awry - must be learning >>>>>>> something with the addition of the new text that it feeds back to the >>>>>>> 1st >>>>>>> and gets it wrong. >>>>>>> >>>>>>> However, I was easily able to get a perfect read on both of them by >>>>>>> performing a simple threshold filter (turning all text dark black) >>>>>>> before >>>>>>> providing to Tesseract. >>>>>>> >>>>>>> On 8 January 2015 at 00:01, Gokcer Gunes <[email protected]> >>>>>>> wrote: >>>>>>> >>>>>>>> First original Image which sended to tesseract >>>>>>>> [image: Satır içi resim 1] >>>>>>>> Second Original Image which sended to tesseract >>>>>>>> >>>>>>>> [image: Satır içi resim 2] >>>>>>>> and outputs are in image which you already saw >>>>>>>> >>>>>>>> 2015-01-08 1:24 GMT+02:00 Gokcer Gunes <[email protected]>: >>>>>>>> >>>>>>>>> yes i will get and upload them asap >>>>>>>>> >>>>>>>>> >>>>>>>>> 2015-01-08 1:24 GMT+02:00 Gokcer Gunes <[email protected]>: >>>>>>>>> >>>>>>>>>> ah no its not noise there is no noise in original img it just >>>>>>>>>> result of crop in paint >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> 2015-01-08 1:21 GMT+02:00 Allistair C <[email protected]>: >>>>>>>>>> >>>>>>>>>>> Ah, I see - interesting :) The 2nd example isn't quite the same >>>>>>>>>>> - it seems to have some noise on the left hand edge. Also, what PSM >>>>>>>>>>> are you >>>>>>>>>>> using? Can you send original size images you send to Tesseract? >>>>>>>>>>> >>>>>>>>>>> On Wednesday, 7 January 2015 23:05:08 UTC, Gokcer Gunes wrote: >>>>>>>>>>>> >>>>>>>>>>>> they are input image and output result ,inputs have same first >>>>>>>>>>>> line but in first output M is correctly converted to M but in 2. >>>>>>>>>>>> output M >>>>>>>>>>>> is converted to H >>>>>>>>>>>> >>>>>>>>>>>> 2015-01-08 1:00 GMT+02:00 Allistair <[email protected]>: >>>>>>>>>>>> >>>>>>>>>>>>> I see the pictures, yes, but I don't understand the question. >>>>>>>>>>>>> Can you elaborate in detail what the issue is? >>>>>>>>>>>>> >>>>>>>>>>>>> On 7 January 2015 at 22:48, Gokcer Gunes <[email protected]> >>>>>>>>>>>>> wrote: >>>>>>>>>>>>> >>>>>>>>>>>>>> i uploaded them as pictures >>>>>>>>>>>>>> >>>>>>>>>>>>>> 2015-01-08 0:29 GMT+02:00 Gokcer Gunes <[email protected]>: >>>>>>>>>>>>>> >>>>>>>>>>>>>>> yeah resul pictures are in message you cant see them? >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> 2015-01-07 23:56 GMT+02:00 Allistair C <[email protected]>: >>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> Your question is not self-evident, what are you trying to >>>>>>>>>>>>>>>> ask? Can you present your OCR results for each test you are >>>>>>>>>>>>>>>> conducting? >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> On Monday, 5 January 2015 18:16:12 UTC, Gokcer Gunes wrote: >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> <https://lh3.googleusercontent.com/-QtKcTsT8fGY/VKrU2Bz4zZI/AAAAAAAAAD4/nCbru06vKac/s1600/testry2.png> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> <https://lh6.googleusercontent.com/-FH8hQKrguns/VKrU0KDbMjI/AAAAAAAAADw/OMBj2VRxtyo/s1600/testry1.png> >>>>>>>>>>>>>>>>> Examples is at above,what i am doing wrong? >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> -- >>>>>>>>>>>>>>>> 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/9f45174d- >>>>>>>>>>>>>>>> c6ed-4590-ab2c-8e93afb57a26%40googlegroups.com >>>>>>>>>>>>>>>> <https://groups.google.com/d/msgid/tesseract-ocr/9f45174d-c6ed-4590-ab2c-8e93afb57a26%40googlegroups.com?utm_medium=email&utm_source=footer> >>>>>>>>>>>>>>>> . >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>> -- >>>>>>>>>>>>>> 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/ >>>>>>>>>>>>>> CAOyRQ0crYYn0Dkgc%2BrzQU7iHf4jfSUt3D8_ >>>>>>>>>>>>>> FnZq2FYUXOR1LsA%40mail.gmail.com >>>>>>>>>>>>>> <https://groups.google.com/d/msgid/tesseract-ocr/CAOyRQ0crYYn0Dkgc%2BrzQU7iHf4jfSUt3D8_FnZq2FYUXOR1LsA%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>>>>>>>>>>> . >>>>>>>>>>>>>> >>>>>>>>>>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> -- >>>>>>>>>>>>> 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/ >>>>>>>>>>>>> CAORW5vhvxni3pq4VyBPLvnF0Uq94seHyUrKPze7RZLTUz5FzUg%40mail. >>>>>>>>>>>>> gmail.com >>>>>>>>>>>>> <https://groups.google.com/d/msgid/tesseract-ocr/CAORW5vhvxni3pq4VyBPLvnF0Uq94seHyUrKPze7RZLTUz5FzUg%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>>>>>>>>>> . >>>>>>>>>>>>> >>>>>>>>>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> -- >>>>>>>>>>> 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/87df217e-3b26-42bc-8cb8-48597edaca62%40googlegroups.com >>>>>>>>>>> <https://groups.google.com/d/msgid/tesseract-ocr/87df217e-3b26-42bc-8cb8-48597edaca62%40googlegroups.com?utm_medium=email&utm_source=footer> >>>>>>>>>>> . >>>>>>>>>>> >>>>>>>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>> -- >>>>>>>> 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/CAOyRQ0c70UZji0DbybnXHdBXAF8C619odsT3hrMNTgD-DcceNQ%40mail.gmail.com >>>>>>>> <https://groups.google.com/d/msgid/tesseract-ocr/CAOyRQ0c70UZji0DbybnXHdBXAF8C619odsT3hrMNTgD-DcceNQ%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>>>>> . >>>>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>>>> >>>>>>> >>>>>>> -- >>>>>>> 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/CAORW5vh6MZUnUA3txczpaZ2UumLc%3DUHBP%2BMVAF6AVN6ZEe4t4g%40mail.gmail.com >>>>>>> <https://groups.google.com/d/msgid/tesseract-ocr/CAORW5vh6MZUnUA3txczpaZ2UumLc%3DUHBP%2BMVAF6AVN6ZEe4t4g%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>>>> . >>>>>>> >>>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>>> >>>>>> >>>>>> -- >>>>>> 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/CAOyRQ0euoR8Bf%3D7XiH6r1VBZ4879m%3DM7BRPRw%2B%3D6DM%3DvDMLiXw%40mail.gmail.com >>>>>> <https://groups.google.com/d/msgid/tesseract-ocr/CAOyRQ0euoR8Bf%3D7XiH6r1VBZ4879m%3DM7BRPRw%2B%3D6DM%3DvDMLiXw%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>>> . >>>>>> >>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>> >>>>> >>>>> >>>> -- >>>> 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/CAORW5vi39huYMiNrST2YrbaFQsFWY6CZdy9xdTKY8Ded-SzVWQ%40mail.gmail.com >>>> <https://groups.google.com/d/msgid/tesseract-ocr/CAORW5vi39huYMiNrST2YrbaFQsFWY6CZdy9xdTKY8Ded-SzVWQ%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>> . >>>> >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> >>> -- >>> 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/CAOyRQ0doGbcLyZm-8pcT%2BC%2BAn0a1veGh-a1nvbmxD6Jiw02%2Byw%40mail.gmail.com >>> <https://groups.google.com/d/msgid/tesseract-ocr/CAOyRQ0doGbcLyZm-8pcT%2BC%2BAn0a1veGh-a1nvbmxD6Jiw02%2Byw%40mail.gmail.com?utm_medium=email&utm_source=footer> >>> . >>> >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- >> 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/CAORW5vgbTWcEFRWr2GUa1haxhtuN%3Duw8d%3DrEOmu1UYGUfphXJg%40mail.gmail.com >> <https://groups.google.com/d/msgid/tesseract-ocr/CAORW5vgbTWcEFRWr2GUa1haxhtuN%3Duw8d%3DrEOmu1UYGUfphXJg%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> >> For more options, visit https://groups.google.com/d/optout. >> > > -- > 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/CAOyRQ0eKo_vnJL_S8D4G2K4%2Bm%2BP_s9%3DnuDqMd6NJiP2kh4nmgg%40mail.gmail.com > <https://groups.google.com/d/msgid/tesseract-ocr/CAOyRQ0eKo_vnJL_S8D4G2K4%2Bm%2BP_s9%3DnuDqMd6NJiP2kh4nmgg%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > For more options, visit https://groups.google.com/d/optout. > -- 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/CAORW5viktkxg9TfBknQ2DPvvUY0qfkguas%2B%2Bb6pBpq4r8gsGkQ%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
