What if We Only Use Real Datasets for Scene Text Recognition? Toward Scene Text Recognition With Fewer Labels.

Jeonghun Baek, Yusuke Matsui, Kiyoharu Aizawa

Browse the full CVPR paper archive.