When it comes to generating images automatically based on your own ideas, the two most popular artificial intelligences among users are DALL-E 2 and Midjourney. Both are tools capable of creating images with astonishing quality, to the point where they can appear real.
Now, a heavyweight in the technology industry wants to compete head-to-head with these popular AIs and has launched its own image-generating artificial intelligence. Nvidia, the company known for creating graphics cards for PCs, has developed Perfusion, an AI that allows the creation of images from natural language descriptions.
By simply inputting the image we want to generate using text, Perfusion will produce an image that closely matches the given instructions. Yes, the result will be “approximate” as it is not a perfect image-generating AI (currently, none are).
Despite Perfusion not being perfect (yet), the results obtained through this AI seem to be more accurate than those of its competitors. Nvidia’s artificial intelligence is based on Stable Diffusion, a language model created by LMU Munich and Runway, but it includes certain mechanisms that simultaneously generate multiple concepts in new images.
Perfusion offers users the ability to combine different customized elements with a set of images that function as concepts, thus creating images with much greater precision than other image generators by having greater control in the generation process.
Nvidia is fully immersed in the world of AI. In addition to Perfusion, Nvidia has also developed Omniverse Audio2Face, a tool that enables the generation of 3D animations from audio. Furthermore, it has been known since the beginning of the year that the company is developing AI-driven drivers to optimize the performance of its graphics cards.
Some of the links added in the article are part of affiliate campaigns and may represent benefits for Softonic.
