Extracting Outlines by LayoutLMv3

[tomanick]

Describe Model I am using LayoutLMV3:

I am working on a project that involves extracting outlines from PDF files. Some of the PDF files I am handling do not contain a structured "Contents" or "Outline" chapter at the beginning, making it challenging to extract the document’s structure through traditional methods.

I am considering using LayoutLMv3 to extract the document's outline by analyzing each page.

Could you please confirm whether LayoutLMv3 is capable of performing such a task? Additionally, I would appreciate it if you could provide guidance or example code on how to implement this. Furthermore, the input data would be PDF files or images?

Many thanks!