NVIDIA’s AI researchers have developed a process that can transform a collection of still 2D images into a digital 3D scene near instantly. Called Instant NeRF, or basically inverse rendering, it uses AI to determine how light behaves in the real world, allowing researchers to reconstruct a 3D scene from a handful of 2D images taken at various angles. This process requires only a few seconds to train on a few dozen still photos before rendering the resulting 3D scene within tens of milliseconds.
What is a NeRF? It’s essentially neural networks that represent and render realistic 3D scenes based on an input collection of 2D images. The neural networks are able to fill in the blanks, while reconstructing the scene by predicting the color of light radiating in any direction, from any point in 3D space. Typically, it takes hours to train AI, but Instant NeRF cuts rendering time by using a technique developed by NVIDIA called multi-resolution hash grid encoding, which achieves high-quality results using a tiny neural network that runs rapidly.
- The IdeaPad 3 17-inch laptop is the ultimate notebook for performance at home and beyond, built with an AMD Ryzen 5 5500U Processor, 8GB DDR4 RAM, and...
- Expect more from your entertainment on the massive 17-inch HDplus display with AMD Integrated Radeon 7 Graphics. Dolby Audio delivers crystal-clear...
- A narrow bezel on 4 sides makes for a clean design and larger display, giving you more viewing area and less clutter, while the Ideapad 3's targeted...
If traditional 3D representations like polygonal meshes are akin to vector images, NeRFs are like bitmap images: they densely capture the way light radiates from an object or within a scene. In that sense, Instant NeRF could be as important to 3D as digital cameras and JPEG compression have been to 2D photography — vastly increasing the speed, ease and reach of 3D capture and sharing,” said David Luebke, vice president for graphics research at NVIDIA.
