SceneTok: A Compressed, Diffusable Token Space for 3D Scenes

SceneTok introduces a novel 3D scene tokenizer that compresses view sets into permutation-invariant tokens, achieving 1-3 orders of magnitude better compression than existing methods while maintaining state-of-the-art reconstruction quality. The system enables efficient 3D scene generation in 5 seconds using a lightweight decoder that can render novel viewpoints.

• →SceneTok encodes 3D scenes into compressed, unstructured tokens that are disentangled from spatial grids.
• →The compression achieves 1-3 orders of magnitude improvement over existing 3D scene representations.
• →The system maintains state-of-the-art reconstruction quality despite heavy compression.
• →Novel view rendering is possible from trajectories that deviate from input trajectories.
• →3D scene generation completes in 5 seconds with superior quality-speed trade-offs.