NVIDIA demonstrated its new Instant NeRF technology that pushes the boundaries of the concept of photography. The technology brings together photos taken from different angles in a single scene, creating a three-dimensional virtual scene.
NVIDIA, one of the leading technology companies, supported by artificial intelligence continues to develop graphics technologies. The company, which introduced many technologies such as the technology that turns your doodles into a realistic scene as if you were drawing from Paint, this time showed its technology that brings a new definition to photography.
NVIDIA’s newly developed ‘Instant NeRF’ With its technology called two-dimensional photos only in seconds paved the way for making it completely three-dimensional. Moreover, the concept of three dimensions here is not created as planar like fake three-dimensional photographs, but as 360-degree viewable.
Photos become three-dimensional in minutes:
NVIDIA’s groundbreaking new technology, Instant NeRF, transforms the subject into a three-dimensional object by examining photos taken from different angles with artificial intelligence. As seen in the video above, not only the subject, the subject’s environment transforms into a completely three-dimensional environment.
This technology developed by NVIDIA is not actually the first technology that enables two-dimensional photos to be turned into a three-dimensional scene. These up-to-date technologies called NeRF can also do the same. But the difference between the new Instant NeRF and NeRF is revealed in the completion time of this transaction. Standard technologies include analyzing the photo and rendering the scene. in a few hours Instant NeRF reduces this time to just a few minutes.
In these few minutes, Instant NeRF analyzes photos while calculating angles, lighting, possible color changes in all directions, and many more factors that are not included in the photos. Technology will bring many advantages such as making architectural designs in virtual worlds and enabling autonomous vehicles and robots to make more accurate measurements of real world objects.
Since the technology makes use of NVIDIA’s CUDA Toolkit and CUDA Neural Networks, it can be run with an NVIDIA graphics card. Especially on NVIDIA graphics cards with Tensor cores, the process is even faster.
Those who want to experience the technology can reach the official GitHub page by clicking this link.
Source :
https://blogs.nvidia.com/blog/2022/03/25/instant-nerf-research-3d-ai/?ncid=so-yout-933463-vt37#cid=gtcs22_so-yout_en-us