DocAligner: Automating the annotation of photographed documents through real-virtual alignment

Jiaxin Zhang, Peirong Zhang, Huiyi Cheng, Xinhong Chen, Haowei Xu, Kai Ding, Lianwen Jin

Abstract: Highlights•Achieved automated annotation of real data for various photographed document tasks.•The first to propose document alignment task for document AI.•Proposed innovative model designs tailored for document alignment tasks.•Addressed the lack of training data and generalization issues in document alignment.