csbowei/ART

<div align='center'> <h1 align="center">[ECCV 2026] Anchoring on Reality: Breaking the Pseudo-Target Ceiling in Makeup Transfer</h1> Bo Wei^1*&emsp; <a href='https://scholar.google.com/citations?user=wLTXeNwAAAAJ&hl=en&oi=ao' target='blank'>Xianhui Lin</a>^2†&emsp; Yi Dong²&emsp; Zhongzhong Li²&emsp; Zonghui Li²&emsp; <a href='https://scholar.google.com/citations?hl=en&user=BhmLztgAAAAJ' target='blank'>Zirui Wang</a>²&emsp; </div>

<div align='center'> <a href='https://scholar.google.com/citations?hl=en&user=12MzNVkAAAAJ' target='blank'>Jiachen Yang</a>²&emsp; Xing Liu²&emsp; Hong Gu²&emsp; <a href='https://scholar.google.com/citations?hl=en&user=tmTvoUAAAAJ' target='blank'>Xiaoming Li</a>^3✉&emsp; <a href='https://scholar.google.com/citations?hl=en&user=rUOpCEYAAAAJ' target='blank'>Wangmeng Zuo</a>^1✉ </div>

<div align='center'> ¹Harbin Institute of Technology&emsp; ²vivo BlueImage Lab&emsp; ³Nanjing University </div>

<div align='center'> <small>^* Work done during an internship at vivo.</small>&emsp; <small>^† Project lead</small> </div>

<div align="center"> <p> <a href="https://arxiv.org/abs/2606.31089" target="blank"><img src="https://img.shields.io/badge/arXiv-ART-red" alt="arXiv link"></a>&nbsp; <a href="https://csbowei.github.io/ART/" target="blank"><img src="https://img.shields.io/badge/Project-Homepage-green" alt="project homepage"></a>&nbsp; <a href="https://github.com/csbowei/ART" target="blank"><img src="https://img.shields.io/badge/GitHub-Code-blue?logo=github" alt="GitHub code"></a>&nbsp; <a href="https://huggingface.co/spaces/csbowei/ART-Makeup-Transfer" target="blank"><img src="https://img.shields.io/badge/🤗 Hugging Face-Space-orange" alt="HF Space"></a>&nbsp; <a href="https://drive.google.com/drive/folders/1UfUUAk86fWWzYsZcGOzpTmAWa32IwK21?usp=drivelink" target="blank"><img src="https://img.shields.io/badge/Dataset-MF2K-purple?logo=googledrive&logoColor=white" alt="MF2K Dataset"></a> </p> </div>

This repository provides the model weights for ART: Anchoring on Reality: Breaking the Pseudo-Target Ceiling in Makeup Transfer.

ART is a makeup transfer framework that applies a reference cosmetic style to a source face while preserving identity, geometry, and fine-grained makeup details.

Files

Higher-resolution checkpoints are not planned for the current release.

Usage

These LoRA adapters are for `black-forest-labs/FLUX.1-Kontext-dev`.

Refer to the project repository: https://github.com/csbowei/ART

Example:

# Makeup transfer: apply the reference's makeup onto the input face
python infer.py \
    --task        mt \
    --input       ./examples/source/src_001.jpg \
    --ref         ./examples/ref/ref_001.jpg \
    --lora_path   /path/to/art_transfer_lora_1024.safetensors \
    --resolution  1024

# Makeup removal: strip cosmetics from the input face
python infer.py \
    --task        demakeup \
    --input       ./examples/ref/ref_002.jpg \
    --lora_path   /path/to/art_demakeup_lora_1024.safetensors \
    --resolution  1024

Citation

@article{wei2026art,
  title={Anchoring on Reality: Breaking the Pseudo-Target Ceiling in Makeup Transfer},
  author={Wei, Bo and Lin, Xianhui and Dong, Yi and Li, Zhongzhong and Li, Zonghui and Wang, Zirui and Yang, Jiachen and Liu, Xing and Gu, Hong and Li, Xiaoming and Zuo, Wangmeng},
  journal={arXiv preprint arXiv:2606.31089},
  year={2026}
}