simple-leftrefill-inpainting 0.0.1

This is a wrapper api of paper LeftRefill

Simple-lama-inpainting

Simple pip package originally from LeftRefill[1] inpainting with reference. Edited from the wrapper from StreetUnveiler[2]

The API design is similar to simple-lama-inpainting.

Installation

pip install simple-leftrefill-inpainting

CLI

simple_leftrefill <path_to_inpainted_image> <path_to_mask_image> <path_to_reference_image> <path_to_output_image>

How to use in your code

from diffusers.utils import load_image
from simple_leftrefill_inpainting import LeftRefillGuidance

model = LeftRefillGuidance()

result = model.predict(
    load_image("./assets/inpainted.png"),
    load_image("./assets/mask.png"),
    load_image("./assets/reference.png"),
)

result.save("./assets/result.png")

Illustration

Inpainted Input:

Inpainting Mask:

Reference Mask:

Result:

License

Please follow both LeftRefill license and the license in this repo.

Reference

[1] Cao et al. LeftRefill: Filling Right Canvas based on Left Reference through Generalized Text-to-Image Diffusion Model. CVPR 2024

@inproceedings{cao2024leftrefill,
      title={LeftRefill: Filling Right Canvas based on Left Reference through Generalized Text-to-Image Diffusion Model}, 
      author={Chenjie Cao and Yunuo Cai and Qiaole Dong and Yikai Wang and Yanwei Fu},
      booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
      year={2024},
}

[2] Xu et al. 3D StreetUnveiler with Semantic-aware 2DGS - a simple baseline. ICLR 2025

@inproceedings{xu2025streetunveiler,
  author       = {Jingwei Xu and Yikai Wang and Yiqun Zhao and Yanwei Fu and Shenghua Gao},
  title        = {3D StreetUnveiler with Semantic-aware 2DGS - a simple baseline},
  booktitle    = {The International Conference on Learning Representations (ICLR)},
  year         = {2025},
}