AI image generation has come a long way in the last couple of years – and we have a feeling, for better or worse, it has a long way still to go. Every day it seems the tech learns how to do something new and inventive – like turning 2D images into 3D models.
DreamCraft 3D is able to take a flat image created by a text-to-image generator and transform it into an impressively high-quality 3D render. (Looking for more AI art? Check out the weirdest images created by AI.)
As spotted by 80LV, the generation model was shared to YouTube this week, where it is described as "a hierarchical 3D content generation method that produces high-fidelity and coherent 3D objects. We tackle the problem by leveraging a 2D reference image to guide the stages of geometry sculpting and texture boosting." It's particularly impressive how the 3D models manage to create believable versions of the 'backs' of each object (such as, in the examples shown, a dog's tail) complete with accurate textures and proportions.
DreamCraft 3D "leverages a 2D image generated from the text prompt and uses it to guide the stages of geometry sculpting and texture boosting. When sculpting the geometry, the view- conditioned diffusion model provides crucial 3D guidance to ensure geometric consistency," according to the project's GitHub page.
It seems there's no end to what AI can do right now. But of course, questions surrounding ethics and copyright have plagued the tech since it burst onto the scene just over a year ago. We recently talked to one illustrator about the raw impact of AI on artists' lives.