Open dhgrs opened 6 years ago
The method using in this project can be used in syllable segmentation case? What's your suggestion about syllable segmentation?
Hi, sorry for late reply!
In master branch, I do charactor(hiragana/katakana)-based segmentation. And I think Japanese charactors are also syllables in many cases. So this project can be used in syllable-based segmentation.
But the accuracy is too bad to use now. I'll try to upgrade the result. If you have any ideas, please tell me!
FYI: I tried Julius segmentation-kit in julius branch. And the result is bad too.
Any discussions, questions and others are welcome here!