Open wusphinx opened 5 years ago
Thanks for the suggestion, yes, I tried several ocr services before. The current reason by default we are using tesseract is that it is free and easy to access. Would plan for make NoSmoke able to accept different types of third party ocr services.
Currently
nosmoke
usestesseract
as default text recognition, buttesseract
is not good at the recognition of Chinese, do you have any plan of abstractingOCR
module to use other identification tools, such as character recognition baidu-ocr?