Create free-viewpoint portraits from videos
Turn casually captured videos into 3D portraits
Transform casual videos into free-viewpoint portraits
Turn casual videos into 3D portraits
test
Create talking face animations from still images and audio
Turn casual videos into 3D free-viewpoint portraits
Create a talking face video from text
Turn casual videos into 3D portraits
Turn casually captured videos into free-viewpoint portraits
Teste
Turn casual videos into photorealistic 3D portraits
My own space
Nerfies: Deformable Neural Radiance Fields is a cutting-edge tool that transforms a single video into a free-viewpoint, talking 3D portrait. By leveraging Neural Radiance Fields (NeRF) technology, it captures intricate details of a subject's appearance and allows users to view the portrait from any angle. This innovation opens up new possibilities for immersive communication, entertainment, and creative expression.
What hardware is required to run Nerfies?
Nerfies is optimized to run on modern GPUs with sufficient VRAM (at least 8GB recommended). A fast CPU and ample RAM (16GB or more) are also recommended for smooth processing.
Can I use any video format?
While Nerfies supports most common video formats, optimal results are achieved with high-resolution, well-lit videos. Ensure the video is clear and sharply focused for the best 3D reconstruction.
What file formats are exported?
Nerfies typically exports 3D models in standard formats like OBJ or GLB, along with texture maps. Rendered images can be exported as PNG or JPEG files for immediate use in various applications.