Virtual Puppet Project is now on Flathub:
It's a Vtuber app:
With your webcam, this app try to detect your face emotions and position to reproduce them on a 3D model.
The app is provided with a duck as default 3D model, but you can import new one in format .glb, .vrm and .scn.
The app is still in developpement, but I tested it and it worked.
All I miss now is a Vtuber model generator. Something that work like a character designer tool in a RPG video game.
You choose all the physical characteristics of your model:
- Eyes size, shape, position
- Nose size, shape, position
- Mouth size, shape, position
Then you choose the clothes and accesoiries and you export your Vtuber model in directly usable format.
Sadly, I don't have enough knowledge and experience to do it.
This is pretty cool. A few questions:
1. Is it FLOSS?
2. Can you perhaps show a demo somewhere?
3. How hard is it to make new models?
Yes, it is FLOSS. MIT licence. To detect face emotions and position, it use OpenSeeFace in BSD 2 clause licence: https://github.com/emilianavt/OpenSeeFace
Warning: OpenSeeFace use a little bit of C#. It can be a problem for some people.
For the demo, maybe later. But it will not show a lot. You will see a duck that move his head and blink his eyes. I don't have another model.
For making a model, I am not an expert about this.
After a few minutes research, you can make a model with Blender but not export directly in .vrm.
To convert from blender to .vrm format, you can use blend2vrm:
It's in GPLv3 licence.
But I haven't tested it. So, I don't krow how it is.
Also, you couldn't know that, but even on my Linux installs, I generally don't use Flatpak. Due to this, linking to the project's source code is more helpful to me.
Le réseau social de l'avenir : pas de publicité, pas de surveillance institutionnelle, conception éthique et décentralisation ! Gardez le contrôle de vos données avec Mastodon !