Abstract: The present scene text extraction methods generally include two independent steps of text detection and text recognition, which is not beneficial to the interaction of feature information ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results