The AI department of chip manufacturer Nvidia has announced a software called Neuralangelo that uses artificial intelligence to create detailed 3D models of iconic sculptures, buildings or even everyday objects. All that is required is a video shot with a smartphone or drone. Neural surface reconstruction uses software to analyse the input and calculate the object’s 3D shape. An initial 3D representation is created and further refined to reflect intricate details. Nvidia’s researchers have said that Neuralangelo is able to display complex surfaces missed by previous models. The technology utilises Neural Radiance Fields, or NeRFs, a development in the field of artificial intelligence that enables 3D models made with just a few dozen 2D photos.
Applications for Neuralangelo include virtual worlds such as video games, where digital twins of real objects and buildings can be created with comparatively little effort. Industrial designers and robotics companies can use the technology to simulate events in a plant or building on the computer. At the Conference on Computer Vision and Pattern Recognition in Vancouver, Nvidia will present more details of Neuraangelo along with 30 other AI projects. Nvidia is promoting the software as a tool that will help recreate the real world in the digital world. Nvidia has reinvented itself as an “AI Computing Company” rather than simply a graphics chip manufacturer.
