Authors: Chenyu Liu, Fengren Wang, Jiajia Wu, Jinshui Hu, Bing Yin, Cong Liu

Affiliation: iFLYTEK Research

Description: By treating this problem as a retrieval task, we presented DOCument OCR Retrieval (DOCR). Our method consists of three building blocks:
1) layout & document level analysis, within OCR and Key-Value Extracting included;
2) sequence labeling and parsing for each question;
3) a fuzzy search algorithm for retrieving and ranking.