The number of artificial intelligences that transform written texts into visual arts is increasing. As it is known, last month, Google introduced Dreamfusion, which transforms written sentences into three-dimensional objects. Now a similar technology comes from Nvidia. Nvidia researchers today announced Magic3D, a 3D visual generator powered by artificial intelligence.What does Nvidia Magic3D offer?
We can call Magic3D an artificial intelligence that works thanks to machine learning. For example, when you enter text such as “blue poison dart frog sitting on a water lily”, you can get 3D results up to 512 x 512 pixels in about 40 minutes.
Reportedly, Magic3D uses a two-stage system and Nvidia’s Instant NGP to convert text to images. Thus, it takes a rough model created in low resolution with the support of artificial intelligence and converts it to a higher resolution. Nvidia, on the other hand, says its new AI model can produce objects about twice as fast as Google DreamFusion.
In addition, Nvidia reported that models created in Magic 3D can be used in video games or CGI (Computer-generated Imagery) scenes.
21709
