Replies: 1 comment 1 reply
-
Hi @lilouartz 👋, for pure OCR docTR works well as you can see: (tested with We don't have any logic yet to extract only tabular data but you could try to combine: https://huggingface.co/docs/transformers/model_doc/table-transformer with docTR for OCR or if the table is on a fixed position work with the predictor hooks (https://mindee.github.io/doctr/using_doctr/using_models.html#advanced-options) to limit the area Hope this helps :) Best regards, |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
My dataset are photos such as this one.
My use case is to extract the "Supplement Facts" table.
Would doctr be a good choice for such a use case?
Beta Was this translation helpful? Give feedback.
All reactions