Authors: VIS-OCR

Description: An end-to-end text localization result submitted by Baidu VIS-OCR, the former OCR team at Baidu IDL.