Neural Radiance Field
A deep learning technique that trains a neural network to represent a continuous 3D scene from a sparse set of 2D images by predicting light and density.
Think of It Like This
Like a sculptor learning the exact shape of a statue just by looking at five photographs taken from different angles.
NeRFs operate by taking a 3D coordinate and viewing angle as input, and outputting the color and volume density at that specific point. Ray marching is then used to render novel 2D views of the scene. While historically slow to train and render, modern optimizations have made them highly viable for VR, AR, and special effects.