search

microsoft / table-transformer

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.
MIT License
2.01k stars 231 forks source link

How to handle the nesting of tables in PDF? #160

Open akshay-bothra opened 6 months ago

akshay-bothra commented 6 months ago

I have a use-case which as outer table and table inside the cell. How should I handle the scenario?

bsmock commented 6 months ago

Hi, I think with anything in document information extraction, the first question I would ask is: what would you like the final result to be? That question comes before deciding an appropriate way of handling it.

I would also ask: can you share an example?

But regardless, our pre-trained models were not trained on any cases where one table bounding box overlaps with another table bounding box. So if this is your case, you'll likely need to train/fine-tune your own model.

Best, Brandon