Authors: VIS-OCR

Description: An improved end-to-end text localization and recognition result submitted by Baidu VIS-OCR.