Authors: Weidong Xu

Affiliation: China Mobile Cloud Center

Description: First, obtain the ocr information in all test set images through the ocr engine, then obtain the corresponding instance ID of the text line through a structured model, and finally convert it into a submission format