Research
Totem: Object-Centric 3D Memory for World-Consistent Video Generation
arXiv:2610.15807v1 Announce Type: new Abstract: Generating coherent videos grounded in a unified underlying world remains a major challenge, particularly under large viewpoint changes, where existing
arXiv:2610.15807v1 Announce Type: new Abstract: Generating coherent videos grounded in a unified underlying world remains a major challenge, particularly under large viewpoint changes, where existing methods often struggle to maintain stable representations of individual objects. To this end, we introduce Totem, an object-centric 3D memory framework for world-consistent video generation. Totem first represents each selected object as a persistent 3D memory aligned to a shared world frame, assigning it a dedicated state. Based on these memories, we design a dual-stream 3D conditioning architecture that combines object memories with scene geometry for target-view synthesis. Moreover, these object memories constrain the model's attention with 3D structural information, thereby preserving object surfaces and relative spatial layouts. Extensive experiments on world-generation and object-centric benchmarks demonstrate that Totem generates videos depicting a more object-consistent 3D world under challenging camera trajectories, while supporting versatile applications.
Related
- 3AM: 3egment Anything with Geometric Consistency in Videos
- GeoFlow: Enforcing Implicit Geometric Consistency in Video Generation
- Towards Realistic and Consistent Orbital Video Generation via 3D Foundation Priors
- Learning to Retrieve: Internalizing Memory Retrieval for Video World Models
- OctWorld: Long-Range World-Consistent Video Generation with Octree-Based 3D Mapping
Source: arXiv cs.CV | 2026-10-14