The first three dimensions describe spatial geometry, while the fourth represents time. A 4D reconstruction can recover an animated character, moving object or changing scene from images or video rather than producing only one static 3D model.
The method must maintain identity and geometry across frames while separating camera movement from subject motion. Outputs can support animation, visual effects, telepresence and analysis of dynamic behavior.
ELI5
4D reconstruction builds a 3D representation and also records how it changes through time. It is a moving 3D reconstruction rather than a single frozen shape.
For example, video of a walking person can become an animated 3D character that preserves the person's shape and motion. The system must keep body parts consistent even when the camera angle or pose changes.
