Trainning tesseract for a new language from scratch that does not exist in Tesseract

[haru...@gmail.com]

The steps mentioned here for [tessercat 3.0-3.02][ https://github.com/tesseract-ocr/tesseract/wiki/Training-Tesseract-3.00%E2%80%933.02 ] is not clear nor I could find any clear documentation about that:

It is mentioned that the following dataset is required:

    tessdata/eng.config
    tessdata/eng.unicharset
    tessdata/eng.unicharambigs
    tessdata/eng.inttemp
    tessdata/eng.pffmtable
    tessdata/eng.normproto
    tessdata/eng.punc-dawg
    tessdata/eng.word-dawg
    tessdata/eng.number-dawg
    tessdata/eng.freq-dawg


But, didn't explained what are the formats or what they actually are?

The language I am working on is not included in utf-8, but is in utf-16, though it has its official unicode code-point range.

From what I understood so far,

eng.word-dawg : I need to create a text file mylang.txt with one word in each line. Words will in the language in which I am working on and the letters too. And then convert a dawg file. I assume the command for that is

    wordlist2dawg mylang.txt mylang.word-dawg

eng.number-dawg : Create a text file mylangnum.txt with the numerical characters - one in each line (0 to 9). Then covert it to mylang.number-dawg


eng.freq-dawg : Same step as eng.word-dawg file, but with the most frequent words ( frequent words could be retrieved for example after processing a certain dataset like newspaper dataset ) starting with the most frequent word in first line ( no need for frequency) then followed by the next frequent word in second line and so on.

I don't know about the rest of the 7 remaining files.

Could someone please direct me to better tutorial to add a new language in tesseract.

OR. Verify my above assumption and tell me about the remaining 7 files. And how to proceed further after having all the 10 files.

The steps : [tessercat 3.0-3.02][ https://github.com/tesseract-ocr/tesseract/wiki/Training-Tesseract-3.00%E2%80%933.02 ]

Generate Training Images

Make Box Files

Bootstrapping a new character set

Tif/Box pairs provided

is still bit confusing to me.

Working with python on Ubuntu 16.04 LTS, tesseract version 3.04.01 ( installed with sudo apt install tesseract-ocr , and is working perfectly for english language)

I am new in this field, sorry if I made any mistake.

If the requirement is to upgrade the tesseract to version 4 first. Then, do  I need to uninstall the previous pervious version or override with some update command ? ( will the PPA of alex-tesseract 4 will work for overriduing the version?)

Thank you.

[Shree Devi Kumar]

For tesseract 3, and training language similar to vie, take a look at vietocr and jtessboxeditor. 

--
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 tesseract-oc...@googlegroups.com.
To post to this group, send email to tesser...@googlegroups.com.
Visit this group at https://groups.google.com/group/tesseract-ocr.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/6c54f502-0c92-424f-87ca-77fe58694d53%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[haru...@gmail.com]

 Hi, you might have got confused with my other question. I am actually working on two languages. Neither of them are currently present in Tessseract. While one of them has somewhat similar  script/ letters with vie. This one has no connection/ totally different with any of the language currently available in tesseract. And also both the language has no connection with the vie. Thanks.

[Shree Devi Kumar]

jtessboxeditor offers tesseract training for version 3.0x that's why I mentioned it.

For tesseract4, training steps are very different.

On Sat, Mar 30, 2019 at 1:14 PM <haru...@gmail.com> wrote:

 Hi, you might have got confused with my other question. I am actually working on two languages. Neither of them are currently present in Tessseract. While one of them has somewhat similar  script/ letters with vie. This one has no connection/ totally different with any of the language currently available in tesseract. And also both the language has no connection with the vie. Thanks.

--

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 tesseract-oc...@googlegroups.com.
To post to this group, send email to tesser...@googlegroups.com.
Visit this group at https://groups.google.com/group/tesseract-ocr.

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/5024a369-90bc-4303-bc34-fdfeb1e8a19b%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

____________________________________________________________
भजन - कीर्तन - आरती @ http://bhajans.ramparivar.com

[haru...@gmail.com]

Ok. thanks.

Could you guide me on how to train in tesseract 4?

On Saturday, March 30, 2019 at 1:28:19 PM UTC+5:30, shree wrote:

jtessboxeditor offers tesseract training for version 3.0x that's why I mentioned it.

For tesseract4, training steps are very different.

On Sat, Mar 30, 2019 at 1:14 PM <haru...@gmail.com> wrote:

 Hi, you might have got confused with my other question. I am actually working on two languages. Neither of them are currently present in Tessseract. While one of them has somewhat similar  script/ letters with vie. This one has no connection/ totally different with any of the language currently available in tesseract. And also both the language has no connection with the vie. Thanks.

--
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 tesser...@googlegroups.com.

To post to this group, send email to tesser...@googlegroups.com.
Visit this group at https://groups.google.com/group/tesseract-ocr.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/5024a369-90bc-4303-bc34-fdfeb1e8a19b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[shree]

On Sunday, March 31, 2019 at 9:55:23 AM UTC+5:30, haru...@gmail.com wrote:

Ok. thanks.

Could you guide me on how to train in tesseract 4?

 

See https://groups.google.com/forum/#!topic/tesseract-ocr/Rz0J27855hQ

https://github.com/tesseract-ocr/tesseract/wiki/TrainingTesseract-4.00

If you want to use the automated tesstrain.sh method, you will have to add the new language code to tesseract/src/training/language_specific.sh and create langdata (script unicharset, training_text, wordlist etc) for the same.

[haru...@gmail.com]

I will have a look into this. Thanks.