Visual Matching is Enough for Scene Text Retrieval.

Lilong Wen, Yingrong Wang, Dongxiang Zhang, Gang Chen

Browse the full WSDM paper archive.