search

lluisgomez / TextProposals

Implementation of the method proposed in the papers " TextProposals: a Text-specific Selective Search Algorithm for Word Spotting in the Wild" and "Object Proposals for Text Extraction in the Wild" (Gomez & Karatzas), 2016 and 2015 respectively.
https://github.com/lluisgomez/TextProposals
193 stars 82 forks source link

How to re-train weak classifier model #17

Open ndquangr opened 7 years ago

ndquangr commented 7 years ago

Thank you for your excellent work. If it is possible, could you please guide me how to re-train model to use in text/non-text classifier (trained_boost_groups.xml)? I tried to use this with Korean character but the result seems not good enough. Thank you very much!

gachiemchiep commented 7 years ago

Maybe this Link will help you. Same author and he also provided sample data in english. So in your case, relace the train data with you own and you can easily re-train weak classifier model to match your need

ndquangr commented 7 years ago

Oh, thank you very much.