rembg VS DIS

A bit conflicted as the linked video is also linked from the actual rembg repo but it seems way faster and more detailed to just read the readme at that repo first, and maybe use a video if something doesn’t make sense.

There are open source tools like rembg (https://github.com/danielgatis/rembg) which call into pre-trained models.

You can use Rembg extension to remove background automatically: https://github.com/danielgatis/rembg

Hi guys 📷 I am new to ml deployment. Can anyone help about production deployment ? I made fastapi docker app which removes background from image (https://github.com/danielgatis/rembg). It uses u2net segmentation. I tried aws ec2, lambda, googld cloudrun and so far ec2(t2-large) is the fastest but still too slow. Also, it costs way more than I expected. Are there any other solution that I can deploy ml app with as low as possible? Where do you guys mostly deploy ml app?

I am currently using rembg https://github.com/danielgatis/rembg

rembg

A python script with rembg could work

4.) rembg -- backgrounds be gone. If there isn't an extension in Automatic1111 for this yet, there should be (I haven't checked recently). Same deal as midas, you can point it at a folder and zap the backgrounds off of all your images. Useful in combo with midas and imagemagick if, for instance, you want images of an object on a white background (Stable Diffusion training via LoRAs/Dreambooth may not benefit, but other things like GANs prefer that sort of training image). Useful if you want to "compose" a scene and you have images of dispirate objects/people you want in that scene.

Ran into another issue with " LoadLibrary failed with error 126 Here's the solution: https://github.com/danielgatis/rembg/issues/312

Hello, I had some succes using this repo : https://github.com/xuebinqin/DIS

DIS

Rembg only adds to my frustration. It's a half-assed solution providing sub-par results and acting as a band-aid for the fundamental absence of any transparency handling in diffusion models. DIS is better, but it's a pain to set up and still often needs retouching. Trying to remove backgrounds post-generation will always be an uphill battle because the information just isn't there.

also try to look into https://github.com/xuebinqin/DIS cuz it may help to either achieve the result straight away, or to make a dataset for layering lora if each picture takes to much touch up

Also checkout https://github.com/xuebinqin/U-2-Net. They have a new version in this repo: https://github.com/xuebinqin/DIS

If you want some tricks to make this smoother, try masking the wolf upfront in each frame and then only replacing the wolf itself with the pixels from sd on the second frame and onward. Keep the first frame's sd output for the background stable or change it less quickly. I've found good results segmenting with this, but lots of good options on foreground segmentation are available.

For instance, click "open in colab" at the top of this Image Segmentation notebook