autonomousvision/giraffe
A CVPR 2021 paper implementing compositional generative neural feature fields for controllable 3D scene synthesis.
Not currently ranked — collecting fresh signals.
star history
GIRAFFE is a research project that learns to synthesize and control 3D scenes as compositional generative neural feature fields. It extends NeRF-style implicit representations with a Generative Adversarial Network framework to enable disentangled control over object shape, appearance, and camera viewpoint. The model generates novel scenes by composing multiple objects with different attributes into a unified neural radiance field, producing image animations with controllable camera trajectories and object translations.
Frequently asked
- What is autonomousvision/giraffe?
- A CVPR 2021 paper implementing compositional generative neural feature fields for controllable 3D scene synthesis.
- Is giraffe open source?
- Yes — autonomousvision/giraffe is open source, released under the MIT license.
- What language is giraffe written in?
- autonomousvision/giraffe is primarily written in Python.
- How popular is giraffe?
- autonomousvision/giraffe has 1.2k stars on GitHub.
- Where can I find giraffe?
- autonomousvision/giraffe is on GitHub at https://github.com/autonomousvision/giraffe.