Extract table of content of a book with ocr

Question

I'm trying to extract the information of the table of content of a book. Given one or multiple images of a table of content I need to understand the chapters name, the pages, etc. Basically the information of the table of the content of the book.

This is an example of a possible input:

The output should be a dictionary (stupid but structured example):

chapter1: {
  name: "The Basics of Blender"
  page: 1
  content: [
    name: "What ..."
  ]
}

I only found a couple of papers but I did not understand them. Do you know if there are some pertained models for this task?

A simple ocr is not a possible solutions since I need to understand the context maybe with a document layout analysis. There aren't many documentations about this specific topic.

Papers:

TOC structure extraction from OCR-ed books

On the Reading of Tables of Contents