Skip to content

mattdeitke/CVPR2019

Repository files navigation

Updated repository at https://github.com/mattdeitke/CVPR-Accepted-Papers-Viewer


CVPR 2019 Accepted Papers

The main goal of these scripts is to build a page that displays the accepted papers for CVPR 2019 in a way that is easier for humans to parse (see: https://mattdeitke.com/CVPR-2019). Below is an example of what this repository will display, and following that is what CVPR open access currently shows.

In particular, there is functionality to cluster papers based on latent Dirichlet allocation topics, create thumbnail images from the first 8 pages of each PDF, find the abstracts, copy a BibTeX, view the paper and supplementary material, and more. The scripts use Python 3.7 and should work for any past and future CVPR conference (unless they change how they display the pages). Modifications can be made to adapt the format to another conference.

Installation

  1. Clone this repository git clone https://github.com/mattdeitke/CVPR2019

  2. Save the HTML from where the accepted papers are displayed. For CVPR, this year, that would be http://openaccess.thecvf.com/CVPR2019.py.

  3. Install ImageMagick, which can be done using sudo apt-get install imagemagick or using another supported method such as brew install imagemagick.

  4. Run pdftowordcloud.py to generate top words for each paper. The output is saved in topwords.p.

  5. Run pdftothumbs.py to generate tiny thumbnails for all papers. The outputs are saved in thumbs/ folder.

  6. Run scrape.py to generate each paperid, title, authors list by scraping the cvpr2019oar.html page.

  7. Run makecorpus.py to create allpapers.txt file that has all papers (one per row).

  8. Run python lda.py -f allpapers.txt -k 7 --alpha=0.5 --beta=0.5 -i 100 . This will generate a pickle file called ldaphi.p that contains the LDA word distribution matrix. Thanks to this nice LDA code by @shuyo! It requires nltk library and numpy. In this example we are using 7 categories. You would need to change the cvprnice_template.html file a bit if you wanted to try different number of categories.

  9. Generate the abstract files inside abstracts/ folder using getabstracts.py.

  10. Finally, run generatenicelda.py to create the index.html page.

Acknowledgements

Big thanks to @karpathy for his NeurIPS preview and ArXiV Sanity Preserver, which is what this repository builds on! Also a thanks to @tholman for creating a more modern GitHub Corners and @shuyo for the LDA code! Finally, more thanks go to the people at CVPR for openly publishing all of their accepted papers!

Licence

MIT License

Releases

No releases published

Packages

No packages published