Editing Neural Radiance Fields and 3D AI (section)

== <span style="color: #FFFFFF;">Remembering</span> ==
* '''Novel view synthesis''' — Generating photorealistic images of a scene from viewpoints not present in the training images.
* '''Neural Radiance Field (NeRF)''' — An implicit neural representation that maps 3D coordinates + viewing direction to color and density.
* '''Volume rendering''' — Integrating color and density along camera rays through the scene to produce 2D images.
* '''Implicit neural representation''' — A neural network representing a continuous signal (scene, shape, image) as a learned function.
* '''Gaussian Splatting (3DGS)''' — Representing a scene as a collection of 3D Gaussian distributions; renders in real time.
* '''Point cloud''' — A set of 3D points representing the surface of an object or scene; simpler than NeRF.
* '''Camera pose''' — The position and orientation of the camera in 3D space; required input for NeRF.
* '''Structure from Motion (SfM)''' — Computer vision technique estimating 3D structure and camera poses from multiple 2D images (COLMAP).
* '''Multi-view stereo''' — Reconstructing 3D geometry from multiple images with known camera poses.
* '''Text-to-3D''' — Generating 3D objects or scenes from text descriptions using AI.
* '''DreamFusion''' — A 2022 Google paper enabling text-to-3D by distilling knowledge from 2D diffusion models into NeRF.
* '''Instant-NGP (Instant Neural Graphics Primitives)''' — NVIDIA's 2022 acceleration of NeRF using hash-encoded positional embeddings; trains in seconds.
* '''4D NeRF''' — Extending NeRF to dynamic scenes with a time dimension.
* '''Occupancy networks''' — Neural networks predicting whether a point in 3D space is inside or outside a shape.
</div>

<div style="background-color: #006400; color: #FFFFFF; padding: 20px; border-radius: 8px; margin-bottom: 15px;">