Official Repository for ECCV 2024 paper ReLoo: Reconstructing Humans Dressed in Loose Garments from Monocular Video in the Wild.

• Clone this repo: git clone https://github.com/eth-ait/ReLoo
• Create a python virtual environment and activate. conda create -n reloo python=3.10 and conda activate reloo
• Install dependenices. cd ReLoo, pip install -r requirement.txt and cd code; python setup.py develop
• Download SMPL model (1.0.0 for Python 2.7 (10 shape PCs)) and move them to the corresponding places:

mkdir code/lib/smpl/smpl_model/
mv /path/to/smpl/models/basicModel_f_lbs_10_207_0_v1.0.0.pkl code/lib/smpl/smpl_model/SMPL_FEMALE.pkl
mv /path/to/smpl/models/basicmodel_m_lbs_10_207_0_v1.0.0.pkl code/lib/smpl/smpl_model/SMPL_MALE.pkl

You can quickly start trying out ReLoo with a preprocessed demo sequence including the pre-trained checkpoint. This can be downloaded from Google drive. Put this preprocessed demo data under the folder data/ and put the folder checkpoints under outputs/Dance_Game10/.

cd code
bash train.sh

This will launch the trainig from scratch. You can also continue the training by changing the flag is_continue in the model config file code/confs/model/model.yaml. The training usually takes 24-48 hours. The validation results can be found at outputs/.

Run the following command to obtain the final outputs. By default, this loads the latest checkpoint.

cd code
bash test.sh

To test on custom videos, please follow the data structure shown in the data folder of the demo video sequence. The official preprocessing scripts are coming soon.

We have used codes from other great research work, including SDFStudio, VolSDF, NeRF++, SMPL-X, Anim-NeRF, Vid2Avatar and SNARF. We sincerely thank the authors for their awesome work!

Here are more recent related human body reconstruction projects from our team:

• Vid2Avatar-Pro
• MultiPly
• Vid2Avatar
• X-Avatar

If you find our code or paper useful, please cite as

@inproceedings{guo2024reloo,
      title={ReLoo: Reconstructing Humans Dressed in Loose Garments from Monocular Video in the Wild},
      author={Guo, Chen and Jiang, Tianjian and Kaufmann, Manuel and Zheng, Chengwei and Valentin, Julien and Song, Jie and Hilliges, Otmar},    
      booktitle = {European conference on computer vision (ECCV)},
      year      = {2024},
    }