Efficient and Scalable Monocular Human-Object Interaction Motion Reconstruction

📰 News

[2025-12-02] Automatic 4d reconstruction code released!

[2025-12-03] 4DHOISolver code released!

🚀 To Do

[x] Release core inference code.

[ ] Release App

[ ] Release Dataset

🛠️ Installation

# Clone this repository
git clone [https://github.com/username/repo-name.git](https://github.com/username/repo-name.git)
cd repo-name

# Create environment (Modify requirements.txt as needed)
conda create -n es_hmr python=3.9
conda activate es_hmr

# Install PyTorch (Adjust cuda version according to your GPU)
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

# Install other dependencies
pip install -r requirements.txt

# (Optional) Install third-party libraries (e.g., SMPL-X, PyTorch3D)
# pip install "git+[https://github.com/facebookresearch/pytorch3d.git](https://github.com/facebookresearch/pytorch3d.git)"

🏃 Auto-Reconstruction / Demo

To run reconstruction on a sample video:

python demo.py \
    --input_path assets/sample_video.mp4 \
    --checkpoint checkpoints/model_sota.pth \
    --output_dir results/demo_output \
    --visualize

🖥️ Annotate app

📖 Citation

If you find this code useful for your research, please consider citing our paper: