Please enable JavaScript.
Coggle requires JavaScript to display documents.
STAGES OF FILM, AI Art Generators, DESIGN PIPELINE, It is possible to…
-
-
-
-
-
Additional points: In many industries, designs have to conform to real world parametres/constraints which AI art generators do not currently understand
Ou tests show the limitations of current-day text-to-image generation models. We see that the models struggle with counting, understanding spatial relationships, knowledge of shapes, and understanding the physics of the world such as how full an object is.
these models seem to do a good job of capturing the meaning of words that have an associated pictorial representation i.e. objects that can be seen in the physical world. However, the models aren’t as strong as understanding words that aren’t objects.
AI art generators also currently lack an understanding of spatial relationships, knowledge of shapes, counting, language/context and real world physics.