-
Notifications
You must be signed in to change notification settings - Fork 888
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ful #154
Comments
To make a dictionary, for tesseract, is still too complicated tessercat should be arranged in the way, that it's doing that automatically, on the base of a file containing the words, of the language, optionally followed by the meaning in one or more other languages , or by example frases (see below) Please fix / program that in the tesseract program, so that it includes a dictionary , from an input file of words, or parts of words, like the following. In a future version of tesseract, one should include in the ocr that one can make a --dictionary option so that, ocr-ing a dictionary, the program itself re-aranges that into such a word-list / dictionary fiole which can be input for a better traineddata file (inclusive dictionary) for tessercat itself wulundu Katze/deu |
Is this the same as Pulaar language? Or is it the Fula language? There seem to exist different scripts for that (based on Latin, Arabic or other scripts). |
The data which you provided is not sufficient for the current Tesseract, but made for the old Tesseract 3 recognizer. There is also no training text. The included word lists are empty. |
ful.traineddata.zip I trained that now with tessercat 5 . the file ful.traineddata.zip is not zipeed, one can remote the .zip behind . The files used to provide that, are in the ful.daten.zip file. I didnt make yet a dictionary. That language is extended over a big area and has many names, such as tukulor , peulle, pulaar, fula, fulfulde, bolle fulbe , ... |
I want to know, when I producing new box files of aonther text, for training tesseract, if after the first run with any data the relevant informations from these data are already "embedded" into the traineddata file, so that for further training i don't need to use these data again (but only new data) , or if I have to let and use all data accumulated in the folder and add the next data. And, in the first case, if later one want to "remove" sama training data which one added before, if this is possible. With thess 5, I have the problem that after accumulating plenty data / box files of new texts, and run everything togehter, then during trainig the program crashs with any matrix error. If I add only new data (nut including the previously included data again) then the problem dont occure. |
Enclosed an updated trained data file |
ful.frequent_words_list.zip Hier weitere box und jpg files , Wortlisten, und ein neues ful.traineddata für die Sprache Fula / Tukölör / Pulaar Außer einem haben die Dateien kein zip Format, sie wurden nur umbenannt als .zip weil sonst das Hochladen nicht geht (d.h. Umbenennen ohne .zip) |
Ich tue jetzt alle weiteren Dateien und Verbesserungen von ful.traineddata nach : |
ful.traineddata.txt
ful.zip
For the tukulor or ful language, with estimated (about) 80 million speakers , this is a ful.traineddata (pls remove the .txt) produced with tessercat 4 from the enclosed image and box files , from scratch, with about 1 minute training
When I put this into my /usr/share/tesseract-ocr/4.00/ folder, then it works fine
The text was updated successfully, but these errors were encountered: