What is HairDeck?

HairDeck is an open-source repository designed to generate stylized 3D hair cards and spring-bone animation chains como las herramientas de animación avanzada directly from a single input image. Developed by researchers from Tsinghua University, Carnegie Mellon University, and The Chinese University of Hong Kong, the system represents each hair card using tokens and employs a conditional set diffusion model to synthesize entire hairstyles coherently.

The generated output preserves individual hair strands while supplying per-card UV coordinates and automatic skinning via chain diffusion models. This architecture bridges the gap between static image inputs and dynamic, animation-ready 3D character assets.

How does the architecture work?

The core architecture of HairDeck relies on a card VAE that encodes individual hair components and a set diffusion model that jointly constructs the hairstyle. Furthermore, a chain diffusion model generates the spring bones required to drive the hair movement, supporting seamless export to formats compatible with virtual avatars.

The repository supports Linux and native Windows environments, utilizing Python 3.12, PyTorch 2.8, and CUDA 12.8. It integrates specialized visual models and preprocessing utilities such as DINOv3, SAM 3, and BiRefNet to process character features accurately sin configuraciones complejas en móviles.

Why has HairDeck gone viral?

HairDeck is rapidly gaining traction in the open-source community because it drastically simplifies the traditionally tedious task of hair modeling for 3D artists and developers que buscan análisis visual rápido. By automating both geometry generation and bone rigging from a simple photograph, it opens up new possibilities for rapid avatar creation, game development, and digital content pipelines.

How do I install and configure HairDeck?

To install and run HairDeck locally on Linux, execute the following commands in your terminal:

Make sure to authenticate with Hugging Face and download the required model weights using the provided setup scripts before running the image-to-hair generation pipeline.

  1. Clone the official repository using git clone https://github.com/hyz317/HairDeck.git and enter the directory.
  2. Install the python dependencies with python -m pip install -r requirements.txt and pyrender.
  3. Log in to Hugging Face and run the weights downloader script via python scripts/setup/download_weights.py.