AI
OuroWorld: Bringing Any 3D World Alive as Diverse, Endlessly Looping 3D Cinemagraphs
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 07/10/2026, 20:00:00 UTC
Briefing created: 14/11/2026, 22:16:36 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 09/10/2026, 06:18:24 UTC.
Briefings
OuroWorld is a mask-free framework that converts static 3D Gaussian Splatting scenes into looping 3D cinemagraphs using a vision-language model and Fourier-series deformation fields. It achieves high user satisfaction with 70.8%-99.0% of comparisons on 39 generated scenes, outperforming all baselines in vividness, naturalness, and loop coherence.
Key points
- OuroWorld transforms static 3D Gaussian Splatting scenes into looping 3D cinemagraphs using a vision-language model to infer plausible dynamics.
- The framework uses Inconsistency-Robust Periodic 4DGS, which combines a Fourier-series deformation field for looping with a Grounded Drift Field to handle cross-view inconsistencies.
Context
This research addresses the limitation of 3D world models that generate frozen, non-looping scenes. OuroWorld provides a solution for creating dynamic, endlessly looping 3D cinemagraphs, enabling diverse motion and general deformation beyond fluid-like motion limitations.
Limitations
Evaluation is ground-truth-free and limited to 39 reconstructed and generated scenes. Results are based on user-study comparisons rather than automated metrics, and the framework's performance on unseen data remains unverified.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 14/11/2026, 22:16:36 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeiekwedkr7pldalwyh6dpppped7c3o4gg6in6wsnezlxdkc5dxlyvi