slideextract

slideextract is a program to extract slides from videos. It outputs an image for each detected slide. Slide detection is done by comparing consecutive frames for differences within a specific threshold, to account for video encoding artifacts.

For better results and faster extraction you should specify a comparison region, such as the slide number. From a series of unchanging comparison regions slideextract will pick the last frame in hopes of capturing the final state of the entire slide (animated bullet points, annotations, etc...).

# Select a comparision region via gui
slideextract -g video.mp4 outprefix

Once extraction is completed you can view the extracted slides with your favorite image viewer and delete possible duplicates. This often happens when the video jumps back and forth between slides.

You can then combine the slides to a pdf with a tool like convert:

convert `ls -v outprefix*.png` outprefix.pdf

Download

• slideextract-0.3.1.tar.xz

     MD5: 121392f806775f30c9912643bdc0c5b5
  SHA256: 02662909a6febb93b59d7cc806923cbacc7c67c65722001b27f75102428a5f69

  All releases

• GitHub mirror
• Bitbucket mirror

Dependencies

• opencv 4.x

Build Instructions

make
make install # DESTDIR= PREFIX=/usr/local

License

BSD 2-Clause license, see LICENSE for more details.