Rendered at 16:13:46 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
geooff_ 22 hours ago [-]
I love seeing this. Background removal is such a core task used in so many systems, I group it in with speech to text in terms of importance.
I love seeing maturity getting pushed in this space. Congratulations to you and the team!
woadwarrior01 20 hours ago [-]
Why extend an MIT licensed model's weights (BiRefNet) and release it under a cc-by-nc-4.0 license?
snyy 20 hours ago [-]
This pertains to the larger open source licensing discussions that have been happening (as I'm sure you've seen too).
We've released projects under the MIT license before, most notably https://github.com/feyninc/chonkie. While we're not trying to directly monetize on this work, credit goes a long way and helps in other operations.
Recently, attributed usage is shrinking. To be clear, this is not a shrinkage in actual use of our software, just how many people acknowledge that they rely on it.
cc-by-nc is a protection against that. We've been very honest about our work being on top of BiRefNet as we want to extend the original creators the same courtesy. I have no issues if individuals fork/finetune/or otherwise build on top of any open source projects we release, irrespective of license. At minimum, we want acknowledgment if a company chooses to use our software in production.
woadwarrior01 9 hours ago [-]
> cc-by-nc is a protection against that.
I don’t think so. There are CC licenses that allow for commercial usage. Most companies ban their developers from using anything with an NC or GPL style licence (with an exception for the Linux kernel).
Also, you might want to speak to an IP lawyer about the legality of re-licensing a derivative of an MIT licensed work, without attribution.
not_lain 8 hours ago [-]
Hi, thanks a lot for bringing this to our radar, and i have updated the model card to reflect this change, and i will update the library shortly to enforce this choice of license.
I have provided additional information in the same thread above, and will reach out to the original author shortly with an apology and I will see how it goes.
I would also like to apologise to the community and i will do better in the future.
- Hafedh Hichri (Lain)
woadwarrior01 5 hours ago [-]
Hey, thanks for doing the right thing. There's no need for an apology.
not_lain 9 hours ago [-]
I did reach out to Peng (the original author) before the release about license, citation, paper_url with specific emphasis on the the license part and got the green light from him to release the model under the original MIT license or the apache2.0 license since i have pushed for it.
since this is a mistake from my part not iterating with the team about my comms i sincerely apologize for this and I have updated the model card to reflect this change to use the original MIT license, i will also update the library so that all models using the birefnet architecture will be released under the MIT license shortly.
and thanks a lot for bringing this to my radar.
daemonologist 19 hours ago [-]
Btw if you do need a permissively licensed model for this task, there are lots out there, e.g.:
Yep, nobody wants to adopt models with this license because it opens legal questions most users, whether professional or non-professional, don't want to muddy their work with.
qingcharles 19 hours ago [-]
How does it stack up against Adobe's model? The advantage Adobe has is there are essentially two options: (1) Select Subject (95% of the time will give you the result you want); (2) Select Person (helps in the edge cases).
The issue I have is, what is the subject? For instance, you have a person sat on a couch. You hit Select Subject. Is the subject the person, or the person plus the couch?
With Adobe's when I only want the person, most of the time Select Subject works, and when it doesn't I can pivot to Select Person, but that has its own problem because it won't include props that the subject is holding, only the person and their clothing.
snyy 19 hours ago [-]
Your question touches on excellent points.
> what is the subject?
FeyNoBg is an "automatic" model. It automatically detects foreground elements and segments the image. Most of the time, this includes all foreground elements. As you can see in the freekick example (https://drive.google.com/file/d/1MZkAGLwbhNVOZ0Oi7XvpCfSEu9Q...), the model output includes the ball, Messi, and the Liverpool defenders. In your example, FeyNoBg will segment around the person plus the couch.
> I only want the person [including props]
The alternative to automatic models are prompt models and those serve the exact use case you're describing. These allow you to specify the foreground element to include. Everything else is removed. That's the next step for FeyNoBg, converting it from an automatic to a prompt model. Now, answering your question:
> How does it stack up against Adobe's model?
We're better on automatic background removal. Support for selecting a subset of foreground elements is coming soon
qingcharles 18 hours ago [-]
Thank you for the excellent reply.
janrakete 19 hours ago [-]
I'd be very interested in that, too.
nickludlam 22 hours ago [-]
This looks great! What are the resolution limits? From the GitHub readme I saw mention of 1024x1024, but I was able to process an image 1920x2880 just fine. I didn't see any really obvious scaling artefacts, so is there something clever happening behind the scenes?
snyy 22 hours ago [-]
We resize the opacity mask. That tends to scale better
defmetrix 22 hours ago [-]
This is awesome, I just bookmarked your tool. Ive been looking for something simple to use, since Its not easy to use the segment anything tool anymore. Thanks!
snyy 22 hours ago [-]
Thanks! Feel free to open an issue if you run into any issues with the outputs
I was so impressed with remove.bg in Dec 2018; awesome to see the progress in this area, so cool.
filcuk 4 hours ago [-]
For causal use, I've had best success with freemium photoroom.com (export res limit on free use).
Even better than OP's, but it's been a while since I've compared these.
rahulb0802 19 hours ago [-]
How did you assemble your training dataset, especially since you mention some of the considerations with different training mixes in your controlled eval?
Recapping here: Our first train was on 4,000 images from the MaskFactory dataset alone. This improved some benchmarks but regressed on others. We took this as a sign of narrow datasets causing unintended specialization.
In our next run, we assembled 26.1K images from 10 different datasets. We capped the amount of images that could come from one source, to prevent a single type of example from dominating. This composite set covered several cases like crowded scenes, camouflage, high-res subjects, fine objects like hair, blurred backgrounds, etc. We then shuffled everything together and trained FeyNoBg.
terryXyz 6 hours ago [-]
feynobg vs u2-net?
sudb 22 hours ago [-]
very cool! I've been having a fun time porting non-LLM models to iOS/Android (recently had some success with making Bonsai's image generation model run on Android) - have you attempted to run any of the workflows on a mobile device at all? Mind if I try?
snyy 22 hours ago [-]
Please do try, I would love nothing more!
All I ask is you make a PR to https://github.com/feyninc/nobg with your results. We'd love to see what you make, contribute in any way we can, and share onwards.
Of course, it's impossible to know for sure what was LLM processed or not, but some of your posts (like this one) have been getting classified that way.
snyy 22 hours ago [-]
The 4K cap was a judgement call, we didn't want one source to dominate. The license is cc-by-nc, just added it to the hugging face
I love seeing maturity getting pushed in this space. Congratulations to you and the team!
We've released projects under the MIT license before, most notably https://github.com/feyninc/chonkie. While we're not trying to directly monetize on this work, credit goes a long way and helps in other operations.
Recently, attributed usage is shrinking. To be clear, this is not a shrinkage in actual use of our software, just how many people acknowledge that they rely on it.
cc-by-nc is a protection against that. We've been very honest about our work being on top of BiRefNet as we want to extend the original creators the same courtesy. I have no issues if individuals fork/finetune/or otherwise build on top of any open source projects we release, irrespective of license. At minimum, we want acknowledgment if a company chooses to use our software in production.
I don’t think so. There are CC licenses that allow for commercial usage. Most companies ban their developers from using anything with an NC or GPL style licence (with an exception for the Linux kernel).
Also, you might want to speak to an IP lawyer about the legality of re-licensing a derivative of an MIT licensed work, without attribution.
https://github.com/KupynOrest/s3od
https://github.com/Tennine2077/PDFNet
https://github.com/PramaLLC/BEN
The issue I have is, what is the subject? For instance, you have a person sat on a couch. You hit Select Subject. Is the subject the person, or the person plus the couch?
With Adobe's when I only want the person, most of the time Select Subject works, and when it doesn't I can pivot to Select Person, but that has its own problem because it won't include props that the subject is holding, only the person and their clothing.
> what is the subject?
FeyNoBg is an "automatic" model. It automatically detects foreground elements and segments the image. Most of the time, this includes all foreground elements. As you can see in the freekick example (https://drive.google.com/file/d/1MZkAGLwbhNVOZ0Oi7XvpCfSEu9Q...), the model output includes the ball, Messi, and the Liverpool defenders. In your example, FeyNoBg will segment around the person plus the couch.
> I only want the person [including props]
The alternative to automatic models are prompt models and those serve the exact use case you're describing. These allow you to specify the foreground element to include. Everything else is removed. That's the next step for FeyNoBg, converting it from an automatic to a prompt model. Now, answering your question:
> How does it stack up against Adobe's model?
We're better on automatic background removal. Support for selecting a subset of foreground elements is coming soon
https://github.com/feyninc/nobg/issues
Recapping here: Our first train was on 4,000 images from the MaskFactory dataset alone. This improved some benchmarks but regressed on others. We took this as a sign of narrow datasets causing unintended specialization.
In our next run, we assembled 26.1K images from 10 different datasets. We capped the amount of images that could come from one source, to prevent a single type of example from dominating. This composite set covered several cases like crowded scenes, camouflage, high-res subjects, fine objects like hair, blurred backgrounds, etc. We then shuffled everything together and trained FeyNoBg.
All I ask is you make a PR to https://github.com/feyninc/nobg with your results. We'd love to see what you make, contribute in any way we can, and share onwards.
Of course, it's impossible to know for sure what was LLM processed or not, but some of your posts (like this one) have been getting classified that way.