README.md

metadata

language:
  - en
base_model:
  - briaai/Fibo-Edit
pipeline_tag: image-to-image
library_name: diffusers
extra_gated_description: >-
  Bria AI Model weights are open source for non commercial use only, per the
  provided [license](https://creativecommons.org/licenses/by-nc/4.0/deed.en).
extra_gated_heading: Fill in this form to immediately access the model for non commercial use
extra_gated_fields:
  Name: text
  Email: text
  Company/Org name: text
  Company Website URL: text
  Discord user: text
  I agree to BRIA's Privacy policy, Terms & conditions, and acknowledge Non commercial use to be Personal use / Academy / Non profit (direct or indirect): checkbox
license: other
license_name: bria-fibo-edit
license_link: https://creativecommons.org/licenses/by-nc/4.0/deed.en
tags:
  - art
  - image-to-image
  - background-removal
  - image-segmentation

     

Fibo-Edit-RMBG: Professional Background Removal Powered by Fibo-Edit
A fine-tuned version of Fibo-Edit specialized for high-quality background removal. Built on Fibo's foundation, Fibo-Edit-RMBG excels at complex foregrounds, fine detail preservation, and challenging image types.

🌍 What's Fibo-Edit-RMBG?

Fibo-Edit-RMBG is a specialized fine-tuning of the Fibo-Edit model, optimized specifically for background removal tasks. Fibo-Edit-RMBG leverages the powerful Fibo-Edit architecture to deliver state-of-the-art results.

📄 Technical report coming soon. For architecture details, see FIBO and Fibo-Edit.

🖼️ Visual Examples

🔑 Key Features

• Fine-Structure Alpha Matting: Produces clean, precise alpha mattes with accurate opacity estimation along high-frequency edge regions — including hair strands, netting, wire structures, and similarly intricate boundaries.
• Consistent Subject Masking: Produces more complete and decisive masks across the full extent of the foreground subject, with fewer cases of regions being incorrectly split between foreground and background.
• Foreground-Background Separation in Low-Depth-Cue Scenes: Better handles images where foreground boundaries are not well-defined by depth — such as flat illustrations, graphic compositions, and images with overlaid text.

⚡ Quick Start

Source-Code & Weights

• The model is open source for non-commercial use with this license
• For commercial use Click here.

Installation

Install the required dependencies:

pip install diffusers transformers accelerate torch pillow

Usage

Basic Background Removal

import torch
from diffusers import BriaFiboEditPipeline
from PIL import Image

# Load the Fibo-Edit-RMBG model
pipe = BriaFiboEditPipeline.from_pretrained(
    "briaai/Fibo-Edit-RMBG",
    torch_dtype=torch.bfloat16,
)
pipe.to("cuda")

# Load your image
image = Image.open("photo.jpg").convert("RGB")

# Generate foreground mask - no mask needed!
mask = pipe(
    image=image,
    prompt = {'edit_instruction':'Generate a detailed grayscale alpha matte. Map the opaque foreground to white and the background to black. Produce soft, anti-aliased grayscale gradients at the edges of the subject to represent fine details and transparency.'},
    num_inference_steps=10,
    guidance_scale=1.0
).images[0]

# Convert mask to grayscale and use as alpha channel
alpha = mask.convert("L").resize(image.size)
image.putalpha(alpha)

# Save the result as RGBA image with transparent background
image.save("output.png")

🎯 Use Cases

• Hair & Fine-Structure Subjects: Well-suited for images where accurate alpha matting along hair strands, fur, or similarly complex edges is critical to output quality.
• Illustrations & Graphic Compositions: Handles non-photorealistic content where foreground-background boundaries are not clearly defined by depth.
• Graphic Design & Text Cutouts: Produces strong foreground separation where text or typographic elements are the subject.

⚙️ Model Parameters

Recommended inference parameters:

• num_inference_steps: 10 (balance between quality and speed)
• guidance_scale: 1 (recommended for fast results)
• torch_dtype: bfloat16 (for optimal memory usage and speed)
• seed: Results vary across different seed values — we recommend experimenting with a few seeds to find the best output for your specific input.

🤝 Contributing

If you have questions about this repository, feedback to share, or want to contribute directly, we welcome your issues and pull requests on GitHub. Your contributions help make Fibo-Edit-RMBG better for everyone.

If you're passionate about fundamental research, we're hiring full-time employees (FTEs) and research interns. Don't wait - reach out to us at hr@bria.ai

Citation

@article{gutflaish2025generating,
  title={Generating an Image From 1,000 Words: Enhancing Text-to-Image With Structured Captions},
  author={Gutflaish, Eyal and Kachlon, Eliran and Zisman, Hezi and Hacham, Tal and Sarid, Nimrod and Visheratin, Alexander and Huberman, Saar and Davidi, Gal and Bukchin, Guy and Goldberg, Kfir and others},
  journal={arXiv preprint arXiv:2511.06876},
  year={2025}
}

❤️ Star Fibo-Edit-RMBG on Hugging Face and ⭐ Star Bria AI on GitHub to join the movement for responsible generative AI!