harvardnlp / image-extraction

Extract images from PDFs

Code for extracting a representative image from a PDF file using CV.

This code needs to be run on GPU. We include a colab example.

Setup

bash setup.sh

Running

First add a bunch of PDF files to a directory pdfs/.

Next call,

python run.py pdfs/ pics/

The code will attempt to extract an image for each pdf into the pic directory.

image-extraction's Issues

[Error] No pic found

Hi,

Thanks for this awesome resource. Unfortunately, I'm not able to extract image from the pdf.

The following error is encountered. You can reproduce this with the provided collab notebook ImageCheck.ipynb.
I also had to bump the version for pyyaml to get the bash setup.bash running.

rm: cannot remove 'dataset/*': No such file or directory
rm: cannot remove 'out/*': No such file or directory
No pic found for ../pdfs/q2noHUqMkK.pdf
rm: cannot remove 'out/*': No such file or directory
No pic found for ../pdfs/mllQ3QNNr9d.pdf