Multi-Cell Decoder and Mutual Learning for Table Structure and Character Recognition

arXiv:2404.13268v1 Announce Type: cross
Abstract: Extracting table contents from documents such as scientific papers and financial reports and converting them into a format that can be processed by large language models is an important task in knowledge information processing. End-to-end approaches, which recognize not only table structure but also cell contents, achieved performance comparable to state-of-the-art models using external character recognition systems, and have potential for further improvements. In addition, these models can now recognize long tables with hundreds of …

abstract arxiv character recognition contents cs.cv cs.lg decoder documents financial format information knowledge language language models large language large language models papers processing recognition reports scientific table them type