Under Linux, the tool JPdfBookmarks can be use to add a collection of bookmarks to a PDF book in batch by loading information from a text file, which stores the table of contents and page numbers. To prepare this text file manually, I’ve summarized the following steps according to my practice for your reference.

  1. Take snapshots of the content pages and save them into images by using ImageMagick.

    convert -density 300 book.pdf[${page_no}] content_page.png

    Here ${page_no} should be replaced with the numeric page number of the content page.

  2. Extract text from the image of the content page using the OCR tool Tesseract.

    tesseract content_page.png stdout -l eng >> bookmark_text.txt

    The option stdout tells Tesseract to output the extracted text to standard output and -l eng specifies the OCR language as English. To check the list of languages supported by Tesseract, execute tesseract --list-langs.

    $ tesseract --list-langs
List of available languages (8):
chi_sim
chi_sim_vert
chi_tra
chi_tra_vert
deu
eng
fra
osd
  3. Open the generated text file bookmark_text.txt and perform a manual cleaning and reorganization if needed. It is suggested to do this work in Emacs, where the command regexp-builder can help us construct and verify regular expressions that are used for efficient text matching and replacement.
  4. Open the target PDF file in JPdfBookmarks and load the text file bookmark_text.txt by clicking the menu item Load in the Tools menu.
  5. Check the validity of the PDF page targets associated with those bookmark items and finally save the PDF file.