Author Topic: Stable Diffusion AI integration  (Read 1757 times)

2022-09-08, 01:04:46

karlingermano

  • Active Users
  • **
  • Posts: 15
    • View Profile
Hi All,

In the recent months text to image and image to image AI usage are spreading quickly. As Stable diffusion is open source and will radically change the whole industry I think it would be very cool if chaos or specifically Corona would consider to somehow implement this into the software.

I already saw that someone wrote a plugin for C4D which can interactively feed the viewport or the render to the Ai and based on it's prompt it changes the output.

You can see that here: https://www.reddit.com/r/StableDiffusion/comments/x6cbwp/wrote_a_plugin_that_renders_the_cinema_4d_scene/?utm_medium=android_app&utm_source=share

The video shows the viewport feeding to the AI. Now I would be interested in using this in the VFB with Corona rending from CPU and based on the AI prompt the GPU may interactively do the diffusion steps. It could be as an overlay which you can turn on and off.

This would open up a whole new perspective to convert your render to different styles etc.
Definitely would worth to just consider how to incorporate this new powerful and free asset into the renderer.
As I can see competition will be strong. What do you think?