Table Transformer (TATR) is a deep learning model for extracting tables from unstructured documents (PDFs and images). This is also the official repository for the PubTables-1M dataset and GriTS evaluation metric.
The existing approach described in the paper trains two models, one for each for the tasks: table detection and table structure recognition. Did you also explore performing TSR directly given an input (document) image, instead of using the cropped table provided by TD?
Ignoring the cases where a single image has multiple tables, do you have thoughts on what the pros/cons are for such a model?
@bsmock @rohithpv
The existing approach described in the paper trains two models, one for each for the tasks: table detection and table structure recognition. Did you also explore performing TSR directly given an input (document) image, instead of using the cropped table provided by TD?
Ignoring the cases where a single image has multiple tables, do you have thoughts on what the pros/cons are for such a model?
Thanks!