3d-consistency

OmniShowcase 通过几何驱动的训练数据合成与表面法线条件控制，实现从多视角参考图像生成保持 3D 一致性的全视角主体视频，并发布了 Showcase-4K 数据集与 ShowcaseBench 评测集，达到最先进的主体身份与 3D 一致性表现。

Orbia 提出了首个系统评估交互式视频世界模型 3D 一致性的基准，通过受控探索与回访协议，从输入保持、生成内容持久性和 3D 自洽性三个维度对 27 个模型进行评测。

World Observer introduces a joint actor-observer generation framework for video world models, using panoramic observers to keep tracking out-of-view objects so their state stays consistent when they re-enter the actor's view. The paper also proposes world-space metrics and a real/synthetic benchmark for evaluating out-of-view evolution.

PAIWorld enhances diffusion-transformer world models with geometric awareness and cross-view attention to improve multi-view 3D consistency for robotic manipulation tasks, achieving state-of-the-art results on benchmarks.

WorldOlympiad presents a comprehensive benchmark for evaluating video-based world models across physical faithfulness, geometric consistency, and interaction fidelity, revealing significant gaps in current generative models.

CityRAG introduces a video generative model that produces long, physically grounded, 3D-consistent videos of real-world cities using geo-registered data, enabling realistic navigation and simulation for robotics and autonomous driving.