Congratulations for findings. Even DangAmbg does not have support for UTF8 for Indic scripts whereas it works well for English lang only. This point may also be investigated the cause of failure..
Regarding dawg files agree with your point, Debayan, it would be nice to suggest your revised code for replacement after experiment for Indic scripts at your end and feedback your findings is posted in this forum to enable other forum members to perform similar test in their own lang.and feedback to forum. "I made word_in_dawg function in dict/dawg.cpp execute print_dawg_node for every iteration and it does print correctly for english but prints a load of garbage characters for Indic" Kindly post your dawg function in dict/dawg.cpp for benefit of indic forum members. please also clarfify if the Dawg files are emply what will happens *for english *as well as indic ? On Tue, Nov 17, 2009 at 3:46 AM, Debayan Banerjee <[email protected]>wrote: > > I have been investigating the cause of the dictionary files not > working for indic script. I looked at the code that creates the > dictionary ( > http://code.google.com/p/tesseract-ocr/source/browse/trunk/dict/trie.cpp) > and found that there is no wide character support (wchar_t). It stores > one byte per node, and it ends up storing non-printable characters per > node. Does this cause the dictionary to be useless for Indic script? > I did a very basuc experiment by adding just 2 words in the dawg > files. I OCRed the same 2 words from an image, and still got absurd > results. I made word_in_dawg function in dict/dawg.cpp execute > print_dawg_node for every iteration and it does print correctly for > english, but prints a load of garbage characters for Indic. > So, will changing all the char to wchar_t and making necessary changes > to code of memory alloc solve the problem? > > > -- > Regards, > Debayan Banerjee > > Support Free Software > http://deeproot.in > > > > --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
