Nvidia has introduced a new artificial intelligence technology that is capable of creating a 3D view almost instantly. “A few dozen photos”.
Baptized Immediate neurological radiation fields (Instant NeRF), which is capable of training “A Small Neurological Network” At intervals of a few seconds, you are well aware of the angles of the shots. The resulting model is used to create all the data needed to create the 3D environment of the scene in question. According to Nvidia, only a few calculations are required “Ten Micro Seconds”.
In a YouTube video, the American company demonstrates this new technology, which apparently relies on internal tools Kuda Toolkit Where Small Bay Neural Networks.
Thus, in the middle of a room we see a person being photographed from many sides, and, with the magic of technology, go into the three-dimensional space around him.
Creating 3D scenes from photos is nothing new. The key innovation, according to Nvidia, is the speed of implementation. The calculation is a thousand times faster than conventional technologies and takes hours to achieve the same result. There are many potential applications.
“This technology can be used to train robots and self-driving cars to understand the size and shape of real-world objects by capturing 2D images or video footage.
It can also be used in architecture and entertainment to quickly create digital representations of real contexts that creators can adapt and expand. “Explains Nvidia in a blog post.
This change does not require much computer power. Since the neural network is small, a graphics card will suffice. But there is one more drawback: it is desirable that the scene in which the photo was taken was motionless.
“If there is too much movement during the 2D capture process, the AI-generated 3D display will be blurred”, Warns the company. But will not the future always be the same?
Proof : Nvidia