It’s finally here! 🎊
We open sourced our #vtuber motion capture solution at #GoogleIO!
Our new MediaPipe model predicts 478 face landmarks + 52 blendshapes from your webcam and is compatible with any ARKit rigged avatar! 😺🧵
I remember all the late nights trying to make Vtuber face tracking apps by hand like this. No AI, just pure trial and error.
Gemini 3 is very impressive and what I use nowadays for most things too. It's a different era now.
It’s been so long since my last feature launch at Google 😭. This time they made a fancy promo video.
You’ll be able to create custom stickers right in Gboard using your own photos or text powered by Pixel Studio.
Your Pixel just keeps getting better! This time, with new features to help you stay more connected 👯♀️ and be more creative 🎨
See what’s new with the latest #PixelDrop: https://t.co/dgfwjIP4wt
@CrimsynVT_ Yeah, at the moment, it should work if you add eye tracking blendshapes to your model. That is how the current sample character is rigged. If you have a legacy VRM model, it will use bones to track eyes.
Kalidoface Lab was rewritten from the ground up so not all features have been ported over from Kalidoface. It supports ARKit rigged VRM models.
You’ll be able to fine tune the blendshape tracking and adjust your camera angle/position 📐in app.
https://t.co/FyOTNbKuH2
FINALLY
FINNNAAAALLLLLLYYYYYYY
F🫦I🔥N👑A🌵L🌍L 😈Y
@EmojiKitchen
now exists on the web!!!! Officially!!!!! And all you have to do … is google emoji kitchen🌷
https://t.co/t8jb2OC9Fr
It still feels like a dream.
Yes, I am working on #Vtuber tech at Google AI!
💪😎💪
Our new BlazePose GHUM Holistic model predicts full body and hand joint rotations for 3D avatars. Available soon at https://t.co/X6nWTWdkqa. 🐎💃🎵
Paper and live demo at @CVPR for #CVPR2022.