A NeRF is a deep neural network architecture, that takes a set of photos from multiple camera positions of a given scene as an input and outputs a volume representation of this scene. Volumes allow for many common volume rendering interactions like repositioning the camera from any angle or even lighting effects.
