Neural Radiance Field

An implicit neural representation of a 3D scene, introduced by Mildenhall et al. in 2020, in which a small neural network is trained to map any 3D coordinate and viewing direction to a colour and density value. Rendering is performed by volumetric ray marching through this learned field, producing photorealistic novel views from a handful of posed input images. NeRFs have become a core scene representation in XR, robotics, and cultural-heritage capture. For accessibility, NeRFs matter because they give AI systems a dense, queryable 3D model of an environment that can feed downstream tasks such as scene description, object localisation, and tactile-graphic generation for blind and low-vision users.