Always at the forefront in image processing, Nvidia just unveiled theInstant NeRF (Neural Radiance Fields), an imaging AI that turns 2D images into 3D models! Everything here is based on technology developed two years ago by a group of researchers from UC Berkeley, Google Research and UC San Diego. Roughly described, the process consists of mapping the color and light intensity of different 2D images, then generating data to connect these images from different viewpoints to ultimately obtain a 3D scene.
Nvidia specifies that several dozen shots may be necessary for conversion into 3D models, knowing that the models obtained will of course be usable in rendering engines (Unreal Engine), modeling software (Maya) or computer graphics. Nvidia doesn’t specify how many shots were needed to make the demo above, but the result is truly impressive.