Turn small videos into 3D portraits
Generate Talking avatars from Text-to-Speech
Create talking face animations from still images and audio
Turn casual selfie videos into free-viewpoint portraits
demo
sites
Turn casual videos into free-viewpoint portraits
Generate Talking avatars from Text-to-Speech
bambina con i cappelli ricci
Turn casual videos into free-viewpoint portraits
Avatar con voz
Turn casual videos into 3D portraits from any angle
Turn casual videos into 3D portraits for any viewpoint
Nerfies: Deformable Neural Radiance Fields is an innovative machine learning tool designed to convert short videos into 3D portraits. It leverages advanced neural radiance fields (NeRF) technology to create deformable 3D avatars that can dynamically change their poses and expressions. This tool is particularly useful for generating talking portraits or animatable 3D characters from minimal video input.
• Deformable 3D Reconstruction: Creates dynamic 3D models that can animate and change expressions.
• Ease of Use: Requires only a short video input to generate a detailed 3D avatar.
• Real-Time Animation: Enables real-time control over the generated avatar's movements and expressions.
• Customizable Features: Allows users to tweak the avatar's appearance and animation styles.
• High Fidelity Output: Produces high-quality 3D models with realistic details.
1. What type of input does Nerfies require?
Nerfies requires a short video clip of the subject, preferably with varying expressions and poses for optimal results.
2. Can I customize the output avatar after generation?
Yes, Nerfies allows users to customize the avatar's appearance, expressions, and animation styles during and after the generation process.
3. Is Nerfies compatible with all platforms?
Nerfies is primarily designed for use on modern desktop computers with sufficient GPU capabilities. Mobile support is limited at this time.