‍

‍

The KAPTO Classification model aims to perform the following tasks:

1. Assign an entire category to a given document. This functionality is often referred to as Categorisation.

2. Split the document into pages which belong to the same category, again a simple Categorisation task.

3. Identifies multiple instances of the same category in the document (imagine a list of pdf pages representing multiple invoices – then the goal of the model is to split your original content files into various instances of the same document type, but each of them perfectly separated from the others). This functionality is often referred to as Split-out.

4. Re-assemble pieces of information, diluted in different locations, into a single instance, as to say whenever a single instance of a document is split among multiple pages mixed with other information, the task is to re-assemble the various part in a homogeneous document page set. This functionality is often referred to as Recomposition.

‍