I have an open source video player for esports coaches to review gameplay footage (https://www.vodon.gg/)

Could I use something like this (or a library) to easily recognise enemy players that have shown up in frames? I would love to be able to automatically populate bookmarks of interesting moments in the match.

Video player looks cool! Even though this is based for badminton, i have found this one <https://cvit.iiit.ac.in/research/projects/cvit-projects/badm...> quite interesting for sport analytics projects.

I think interesting moments bookmarking is more of an open-ended problem and would very much depend on the game, but large computer-vision models like CLIP have proved to be really useful in recognizing general-purposes activities. You could sample frames uniformly from a game video, index them with a text-captioning model, then find/curate a subset of those captions based on your definition of interesting, and then use those curated captions to look for those moments in future videos, i think!

Hey, that's a really great idea! Thanks, I'm going to play around with it.

Looks pretty cool! I work for a large content company, we would maybe use this if it had:

- A way to run in a container in linux - Or a REST API we could use to run it.

Yes, we should be able to containerize it easily as we already tested it for virtual linux machine. Is there a way to contact you, in case we add this and more features in the future?

Please post again when the Linux version is out!

Will also be great if you create a command-line version.

Thanks for checking out. We are also working on SDK to make it easier to integrate with existing applications. Linux version is almost ready, just looking at ways to distribute dependencies a bit more easily.

Nixpkg, w/ allowUnfree.

If you try it I'd be interested in learning about your experience, as I'm looking into doing something similar early next year (distribution, not similar to the app itself).

Thanks for suggestion! Haven't worked with Nix ecosystem yet, but definitely going to check it out and see what works in the end.

I personally like AppImage best of the "non-native" app package formats, but ideally you could set up one of the native package build CI pipelines somewhere and automagically build native distro packages from your source tree for the most common three or four native package formats and let the community build more "esoteric" packages for themselves from those possibly?

I don't know why but this made me wonder if you could train a model on eye tracking data from videos then use that to predict what people would look at with new content.

This is awesome. Would be great is i could use this across multiple video files and search in a library of files.

This is the feature I'll be building this winter into my Video Hub App

https://videohubapp.com/ && https://github.com/whyboris/Video-Hub-App

I have a branch that detects faces and extracts vectors, it's just a matter of aggregating them and creating a UI so users can search for videos by face (or by dropping an image with a face).

Still a lot of work pending on user interface side, but this is definitely on the TODO list. We are also working to allow users to search for a lot more categories of objects in the upcoming releases. It has been a lot of work, so thanks for feedback and checking out!

You should not take the risk to publish a demo of your commercial product using non-free video material - it will hurt you, even more, when successful. Read a book about copyright and just use creative commons video material.

Good luck!

try www.muse.ai