Support for videos?

Can we add support for video de-duplication? Not an ML expert, but here are some approaches:

https://github.com/Chinmay26/Near-Duplicate-Video-Detection

For *exact* duplicates, it should be much easier - just take a sample of frames and using the existing duper.