Social EgoMesh Estimation (SEE-ME) WACV '25

The official PyTorch implementation of the 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) paper Social EgoMesh Estimation (SEE-ME). SEE-ME is a project focused on egocentric motion estimation. This repository provides the necessary setup instructions, training/testing pipelines, and key components for running the model.

Installation

Environment Setup

conda create -n seeme python=3.8.18
conda activate seeme
pip install -r requirements.txt

SMPL Model Download

To download and prepare the SMPL model, run:

bash prepare/download_smpl_model.sh

Model Checkpoints

Model Component	Checkpoint Path
Interactee Only	Not yet available
Scene Only	Not yet available
Scene + Interactee	Not yet available

Data

Please refer to EgoBody and GIMO.

Usage

Configuration

• Modify Line 114 in the configuration file to set the conditioning based on the checkpoint model used.
• Modify Line 71 in the configuration file to adjust the number of repetitions for testing.

Train

To train a model refer to train.sh.

Test

To test a model refer to test.sh.

Metrics

We refer to evaluation metrics implemented in MLD.

Visualizations

Refer to render.sh.

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository.
2. Create a feature branch (git checkout -b feature-name)
3. Commit your changes (git commit -m "Add new feature").
4. Push to the branch (git push origin feature-name). 5.Open a pull request.

Citation

@misc{scofano2024socialegomeshestimation,
      title={Social EgoMesh Estimation}, 
      author={Luca Scofano and Alessio Sampieri and Edoardo De Matteis and Indro Spinelli and Fabio Galasso},
      year={2024},
      eprint={2411.04598},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2411.04598}, 
}