New AI Technology from Nvidia Produces High-Resolution 3D Models From Text Prompts

The technology industry is currently dominated by generative artificial intelligence (AI). Nvidia, a US-based GPU manufacturer, intends to expand this emerging industry by adding "additional dimensions." The IT giant thus presented Magic3D, a high-resolution generative AI that generates 3D objects from any provided text instructions. By 2030, analysts predict that this industry will continue to expand and reach a market size of $110 billion. After introducing its Magic3D technology, Nvidia seems to enter this sector. The business claimed that it could produce top-notch 3D textured mesh models from input text prompts. [caption id="" align="aligncenter" width="600"] Image credit- Tech Times[/caption] Additionally, a coarse-to-fine procedure is used to comprehend and duplicate the 3D representation of the intended target material. This approach makes use of both low- and high-resolution diffusion priors. Simply told, Magic3D will give you a 3D mesh model of your text prompt if you want to have any kind of thing, like a black dress. Even more specific and detailed prompts, like a Michael Angelo-style statue of an astronaut or a silver candelabra sitting on a red velvet tablecloth with only candle lit, can result in 3D models, including a car made out of sushi, a peacock on a surfboard, and a silver candelabra with a red velvet tablecloth. [caption id="" align="aligncenter" width="1366"] Image credit- Tech Times[/caption] It should be mentioned that the inspiration for these examples came from the questions offered by Nvidia on their website. It demonstrates how the latest AI technology can produce models from incredibly detailed cues. Additionally, the business asserts that DreamFusion, a text-to-3D technology that employs 2D diffusion, was outperformed by Magic3D in terms of speed and 8x greater resolution supervision when creating 3D content. Also read: Magic3D From Nvidia Can Create 3D Models From Text Nvidia also claimed that users could update an input text prompt to alter a 3D model that had already been created. The developers of the new technology, for instance, showed off this edit feature by switching the mesh model of a little bunny sitting atop a stack of pancakes to a metal rabbit perched atop a stack of broccoli.