NeRF

Neural Radiance Fields (NeRF) is the 2020 radiance field technique from UC Berkeley, Google Research, and UC San Diego that synthesizes novel views of complex 3D scenes from sparse 2D images using a coordinate-based neural network. The original paper established the modern radiance field research program and sparked hundreds of follow-up methods improving training speed, quality, editability, and dynamic scene support. NeRF-based methods remain active across research and select commercial deployments alongside newer radiance field approaches like Gaussian Splatting.