Abstract
We describe an off-line handwritten Korean character recognition module for real-time address reading. Our module has a two-stage recognition structure. The first recognition aims at reducing target classes for the second recognition at high-speed. We introduced a method to remove non-similar candidates with an input from the candidate set returned by the first recognizer. The optimal candidate set obtained after the operation of the removal method is used to be the target class set for the second recognition and hence should include the class membership of input without fail and a possible small number of candidates in consideration of the whole processing time. The result of experiment done with the PE92 database has proven the superiority of the removal method.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
