Abstract: We present a novel approach for disentangling the content of a text image from all aspects of its appearance. The appearance representation we derive can then be applied to new content, for ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback