This repository contains the VAREN horse model describen in the paper : VAREN: Very Accurate and Realistic Equine Network, by Silvia Zuffi, Ylva Mellbin, Ci Li, Markus Hoeschle, Hedvig Kjellström, Senya Polikovsky, Elin Hernlund, and Michael J. Black, CVPR 2024.
Follow the instructions to install pytorch3d Create a conda environment and install the required packages.
conda activate pytorch3d
conda install pip
pip install absl-py
Clone the repository. Registed on the project website. Download the checkpoint 'pred_net_100.pth' and place it in the folder:
varen/varen/code/cachedir/snapshots/varen/
Download the horse smal model 'varen_smal_real_horse.pkl' and place it in the folder:
varen/varen/model/
If you want to run the training code you need the dataset. Download and place the extracted scans 'decimated_clean.tar' and the extracted registrations 'registrations.tar' in the folders:
varen/data/scans/decimated_clean/
varen/data/registrations/
If you want to compute the errors on the testset you need the testset. Download 'testset_inside_shape_space.tar' and 'testset_outside_shape_space.tar' and place the extracted files in the folders:
varen/data/testset_outside_shape_space
varen/data/testset_inside_shape_space
To retrain the model, from the 'varen' directory:
./train.sh
To compute the errors on the testset (by default on the outside shape space data):
./predict.sh
If you found the model or any of the pieces of code useful in this repo, please cite the paper:
@inproceedings{Zuffi:CVPR:2024,
title = {{VAREN}: Very Accurate and Realistic Equine Network},
author = {Zuffi, Silvia and Mellbin, Ylva and Li, Ci and Hoeschle, Markus and Kjellström, Hedvig and Polikovsky, Senya and Hernlund, Elin and Black, Michael J.},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
pages = {},
month = Jun,
year = {2024}
}