I have a document in the form of an image. The content of the document is some information in the form of a table. Most often, these are some scanned documents, that is, black and white. However, sometimes there may be originals, where there may be a blue seal, or a company logo of any color (perhaps this information is important for solving my issue).

The point is that I want to remove all the text from the document and get only a template. An example is below.

Document which I get

I would like to get result

I am not asking for a ready-made solution (at your discretion), I would like to hear your advice, recommendations, links on how this can be implemented (and is it possible at all)