top of page
logo Maxime Trassebot

AI X Ceramic

JUNE, 2024
Ceramic
Research project

Fig 1. Canny Edges. Outlines from which to create the first image.

Fig 2-8. Different stages of image refinement in Stable Diffusion.

Mastering Visual Creation with AI (without modeling): How far can we control the results generated by AI?

To begin to answer this question, I undertook a research exercise by creating a red ceramic vase with no CAD or CGI, just a sketch and text. I first drew a sketch, used Stable Diffusion to generate visuals by combining text and outlines, then refined the details with Automatic1111, and finally transformed the images into videos with Luna. After a few days, I obtained a result close to my initial idea. My next step is to experiment with different components for a more realistic rendering.

Fig 9. Video generated on Luna software from an image and a prompt.

Fig 10-11. AI-generated illustrations of the vase.

bottom of page