vid2slides
is a utility to extract slides from videos. It can extract slides taken from videos posted to YouTube, recorded from Zoom, Google Meet, etc. It integrates OCR functionality to recognize what's inside the slides. You can use it to:
This gif was generated by applying vid2slides to this video.
This PDF was generated for the same video.
This list of timings was generated for the same video
video-to-slides
requires command line utils that only work on Unix (Linux/Mac). Installation via conda
:
cd vid2slides
conda env create -f environment.yml
conda activate vid2slides
cd into the vid2slides
directory, and run:
python vid2slides.py input_vid output_json
The result is an output json file that contains info about slides extract from the video.
Use the resulting json file to create a YouTube playlist, or use one of the two utils for create a PDF, GIF or chapter listing for YouTube:
python slides2pdf.py input_json output_pdf
python slides2gif.py input_json output_gif
python slides2chapter.py input_json
ffmpeg
.