抄録
Attention models have been quite researched, and also effectively used for various object detection tasks. As to real scene text detection, it seems that it is still under investigated. In this paper, we use four art-of-the-state attention models to detect scene text, and give a comparison of the results for each models