Generation strand-based hairstyle from single image.

[Paper] [Project Page] [Video]

1. Install CUDA 11.8

Follow the instructions on https://developer.nvidia.com/cuda-11-8-0-download-archive.

Make sure that

PATH includes <CUDA_DIR>/bin
LD_LIBRARY_PATH includes <CUDA_DIR>/lib64

The environment was tested only with this CUDA version.

2. Clone repo, download checkpoints and preprocessed example data:

git clone --recurse-submodules https://github.com/Vanessik/Im2Haircut
cd Im2Haircut
chmod +x ./install.sh
./install.sh

1. Launch the demo:

bash ./scripts/static.sh

2. Find results in the folder: ./exps_inverse_stage/try/examples/{IMG_NAME}

3. Visualize results using tensorbard:

tensorboard --logdir ./exps_inverse_stage --port 6008

To run the pipeline on your own images, first download the model checkpoints and install all required submodules and environments:

chmod +x ./install_submodules.sh
./install_submodules.sh

Then follow these steps:

1. Place your input images in ./data/new_data/img

2. Compute preprocessing data (masks, orientation maps, depth maps, and cameras):

bash ./scripts/preprocess_any_data.sh

3. Run the main inference:

bash ./scripts/static.sh

💻 Code was tested on NVIDIA A100 GPU.

This code is based on the 3D Gaussian Splatting project. For terms and conditions, please refer to LICENSE_3DGS. The rest of the code is distributed under CC BY-NC-SA 4.0.

If this code is helpful in your project, cite the papers below.

This work is based on the great projects:

• Hairstep calculation direction map;

• PERM basis implementation;

• NeuS learning head signed distance function;

• Deep3DFaceRecon_pytorch camera estimation;

• GaussianHaircut hairstyle rasterization using gaussian splatting;

• ml-depth-pro depth estimation;

• VOODOO3D-official implementation of transformer architecture;

@article{sklyarova2025im2haircut,
    title={Im2Haircut: Single-view Strand-based Hair Reconstruction for Human Avatars},
    author=Sklyarova, Vanessa and Zakharov, Egor and Prinzler, Malte and Becherini, Giorgio and Black, Michael and Thies, Justus},
    journal={ArXiv},
    month={Sep}, 
    year={2025} 
}