metadata

library_name: pytorch
language: en
tags:
  - image-to-image

ROME: Realistic one-shot mesh-based head avatars

Paper | Project Page | Github

Model Description

The ROME models can be used to create a personal avatar from a single image. The resulted meshes can be animated and rendered with photorealistic quality. To render a ROME avatar with pretrained weights the FLAME Model and DECA weights are required.

FLAME Project: https://flame.is.tue.mpg.de/modellicense.html

Citations

@inproceedings{Khakhulin2022ROME,
  author    = {Khakhulin, Taras and Sklyarova,  Vanessa and Lempitsky, Victor and Zakharov, Egor},
  title     = {Realistic One-shot Mesh-based Head Avatars},
  booktitle = {European Conference of Computer vision (ECCV)},
  year      = {2022}
}