What's next after signspell: open-sourcing a gesture recognition framework

Thanks so much for all the support on signspell! πŸ™Œ

One thing building this taught me: almost every gesture/action recognition project ends up rebuilding the same infra from scratch β€” capturing frames, labelling landmarks, exporting clean datasets for training.

So I'm working on open-sourcing that layer as a standalone framework/toolkit, so anyone can go from "I have an idea for a gesture-based app" to a labeled, training-ready dataset without reinventing the pipeline each time.

Still early, but if that's useful to you or your team, I'd love to hear what you'd want out of it. More soon!

109 views

Add a comment

Replies

Best

I appreciate that you are open sourcing the infrastructure instead of just another finished application. Thats often where the biggest long term value gets created for developers building in this space.

Β Yes Sure. Thanks .

Β appreciate you notice

One question I'd have is how extensible you're planning to make it. if developers can plug in their own preprocessing, labeling strategies, or export formats, I could see this fitting into a lot of existing workflows.

Β Yes ofcourse,
I am planning to move in this direction. Let us see.