This repository contains the results of various AI Background Removal methods/models tested by me.
I tested 14 different Github repo implementations
/models
/implementations
/software
of AI Background Removal. I tried to stick to FOSS sources.
Three of my test images used I found Here.
If you only want to see the results visit: results.md.
Also if you are hasty you migh wanna have a look at Model Research/Info.
If you wanna go through the whole readme consider viewing the README.md Here as it is a bit wider which presents some of the tables better.
note: the results I present are not definitive, you could perhaps get better results with tweaking the implementation's settings/prompts/parameters etc.
I tested the following:
- https://github.com/OPHoperHPO/image-background-remove-tool - Jump to the results
- https://github.com/plemeri/transparent-background - Jump to the results
- https://github.com/nadermx/backgroundremover - Jump to the results
- https://github.com/danielgatis/rembg - Jump to the results
- https://github.com/MangoLion/bgbye - Jump to the results
- Pinokio - Jump to the results
- ComfyUI - Jump to the results
- https://github.com/siliconflow/BizyAir - Jump to the results
- https://github.com/cubiq/ComfyUI_essentials - Jump to the results
- https://github.com/melMass/comfy_mtb - Jump to the results
- https://github.com/MoonHugo/ComfyUI-BiRefNet-Hugo - Jump to the results
- https://github.com/lldacing/ComfyUI_BiRefNet_ll - Jump to the results
- https://github.com/Bria-AI/ComfyUI-BRIA-API - Jump to the results
- https://github.com/Sanster/IOPaint - Jump to the results
I know people will come here to find the best background removal model or service, but unfortunately there is no "best" for all images. With different images, different models will be better.
Additionally, each model has pros and cons in terms of the set up process & running it. You will have to try each one yourself to see how it goes. Look at the results posted here and find the model that best lines up with your use-case. You may also want to look at Model Research/Info.
I do invite you though to read the rest of the README.md as it is quite interesting and informative and you might find something you wanted. My testing/evaluations was done on a small dataset and with a basic understanding of the different implementations' scripts so it is not conclusive.
CarveKit | Repo | Implemented Neural Networks
CarveKit works with 4 different Neural Networks (Tracer-B7
, U^2-net
, BASNet
& DeepLabV3
).
CarveKit can also be used here on The official website and on a google colab.
Original | Tracer-B7 | U^2-net | BASNet | DeepLabV3 |
---|---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Transparent Background | Repo
Transparent Background is a background removing tool powered by InSPyReNet (ACCV 2022). It has a simple and easy to use GUI.
Normal Results:
Original | Transparent Background |
---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Reversed Results:
Original | Reversed |
---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
BackgroundRemover | Repo
BackgroundRemover works with 3 different background removal methods (u2netp
, u2net
& u2net_human_seg
).
BackgroundRemover can also be used here on The official website.
Original | u2netp | u2net | u2net_human_seg |
---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Rembg | Repo | Implemented Models
Rembg can work with 15 different Models (u2net
, u2netp
, u2net_human_seg
, u2net_cloth_seg
, silueta
, isnet-general-use
, isnet-anime
, sam
, birefnet-general
, birefnet-general-lite
, birefnet-portrait
, birefnet-dis
, birefnet-hrsod
, birefnet-cod
& birefnet-massive
). I did not include the results for u2net_cloth_seg
because it didn't work for any of my test images but you are welcome to go look at them in this results folder in this repo.
Original | u2net | u2netp | u2net_human_seg | silueta | isnet-general-use | isnet-anime | sam | birefnet-general | birefnet-general-lite | birefnet-portrait | birefnet-dis | birefnet-hrsod | birefnet-cod | birefnet-massive |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
BGBye | Repo
BGBye uses 10 different open source background removal models.
BGBye can also be used here on The official website or on an alternate site.
Original | Bria RMBG1.4 | InSPyReNet | U2Net | Tracer-B7 | BASNet | DeepLabV3 | U2Net Human | Open RMBG | ISNET-DIS | ISNET-Anime |
---|---|---|---|---|---|---|---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Pinokio | Link
Pinokio is a browser that lets you install, run, and manage ANY server application, locally.
BRIA RMBG | Link
Original | BRIA RMBG |
---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
RMBG-2-Studio | Link
Original | RMBG-2-Studio |
---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
ComfyUI | Repo
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
I used quite a few ComfyUI nodes that are used to remove backgrounds.
I have also included my ComfyUI Background Removal Workflow Here, where I use all the background remove nodes and all the available models all at once.
BizyAir | Repo
node name: "Remove Image Background"
Original | Remove Image Background |
---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
ComfyUI Essentials | Repo
node name: "RemBG Session"
Original | u2net | u2netp | u2net_human_seg | silueta | isnet-general-use | isnet-anime | sam |
---|---|---|---|---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
MTB Nodes | Repo
node name: "Image Remove Background Rembg"
Original | Image Remove Background Rembg |
---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
ComfyUI BiRefNet Hugo | Repo
node name: "BiRefNet"
Original | BiRefNet |
---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
ComfyUI BiRefNet ll | Repo
node name: "LoadRembgByBiRefNetModeI"
Original | General | General-Lite | General-Lite-2K | Portrait | Matting | DIS | HRSOD | COD | DIS-TR_TEs | BiRefNet-DIS_ep580 | BiRefNet-ep480 |
---|---|---|---|---|---|---|---|---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
BRIA ComfyUI API Nodes | Repo
node name: "Bria RMBG" & "Bria Remove Foreground" (I found the "Bria Remove Foreground" node to be quite impressive).
Original | RMBG-2.0 | Bria Remove Foreground |
---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
IOPaint | Repo
Image inpainting tool powered by SOTA AI Model. Remove any unwanted object, defect, people from your pictures or erase and replace(powered by stable diffusion) any thing on your pictures.
With IOPaint you have 15 background removal models to choose from (RMBG-1.4
, RMBG-2.0
, u2net
, u2netp
, u2net_human_seg
, u2net_cloth_seg
, silueta
, isnet-general-use
, birefnet-general
, birefnet-general-lite
, birefnet-portrait
, birefnet-dis
, birefnet-hrsod
, birefnet-cod
& birefnet-massive
).
Original | RMBG-1.4 | RMBG-2.0 | u2net | u2netp | u2net_human_seg | silueta | isnet-general-use | birefnet-general | birefnet-general-lite | birefnet-portrait | birefnet-dis | birefnet-hrsod | birefnet-cod | birefnet-massive |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
I will try to present my findings from newest/best models/implementations to oldest/worst. You can then Ctrl+F
> model_name
to find everywhere, where that model is used.
RMBG-2.0 | source
RMBG-1.4 | source
Open RMBG / ormbg / Open Remove Background Model | source
BASNet | source
BiRefNet DIS ep580 | source
BiRefNet ep480 | source
BiRefNet Matting | source
BiRefNet General Lite 2K | source
BiRefNet General Lite | source
BiRefNet General | source
BiRefNet HRSOD | source
BiRefNet COD | source
InSPyReNet | source
silueta | source
ISNET anime | source
u2net | source
u2netp | source
u2net_human_seg | source
u2net_cloth_seg | source
sam | source
Creative Commons Attribution Share Alike 4.0 International (CC-BY-SA-4.0)
AI Background Removal Results © 2025 by Courage (Courage-1984) is licensed under CC-BY-SA-4.0