Support for Extracting and Integrating Slides from Video Recordings
Opened this issue · 0 comments
Currently, we do not support adding supportive material like slides or images to the transcripts, but we want to include this feature in the future (see bitcointranscripts/bitcointranscripts#326).
Problem
When it comes to adding slides, the main challenge is obtaining them. Sometimes, the source might link to the slides or provide them, but often, the only available material is a video recording that shows the slides while the speaker is talking.
Solution
We can address this by extracting slides from the video recordings as part of our post-processing workflow. The final output should be timestamped slides that can be easily added to the markdown file of the existing transcript.
By implementing this, we can enhance the quality and completeness of our transcripts, making them more valuable and informative for users.
Existing Tools
There are several tools available that can perform this task. Below are a few (some outdated) resources that I found as part of my research on existing solutions:
- https://github.com/tvdaal/slide-extractor
- https://github.com/spazewalker/slide_extractor
- https://github.com/Slide-extractor-beta/slide-extractor
- https://github.com/szanni/slideextract
- https://pypi.org/project/slide-extractor/ *
- https://github.com/patrickmineault/vid2slides
- https://github.com/derbengale/slide-extractor-from-video
- https://github.com/realChesta/slide-extractor
- https://github.com/radiantly/slide-extractor
- https://github.com/johan456789/slide-extractor
- https://www.youtube.com/watch?v=tzyNRR9bJV8
- https://www.youtube.com/watch?v=Q0BIPYLoSBs (src)