Improving text recognition by distinguishing scene and overlay text.

Bernhard Quehl, Haojin Yang, Harald Sack

Browse the full ICMV paper archive.