top | item 46343260

Why Can't Gemini Generate Images with Transparent Backgrounds?

1 points| rukshn | 2 months ago |ruky.me

2 comments

order

vunderba|2 months ago

It's a pretty well known problem in the generative image domain. GPT-image-1 (and gpt-image-1.5) are outliers in their ability to generate images with a proper alpha channel.

You can still use other models (Flux, SDXL, NB, etc), you'll just need to request that the image be generated on a solid colored blackground. Then just run it through Rembg [1] or similar background removal tool.

[1] - https://github.com/danielgatis/rembg

rukshn|2 months ago

But it's very interesting that ChatGPT can generate with transparent backgrounds where Gemini continues to fails despite being considered as a better image generation model.

The removing background manually not the best approach because it leaves some rough edges and not as smooth as as generating an image with transparent background